INTERACT FORUM

Please login or register.

Login with username, password and session length
Advanced search  
Pages: 1 ... 3 4 5 6 [7] 8 9   Go Down

Author Topic: JRMark -- JRiver's Benchmark Tool  (Read 225986 times)

Matt

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 42034
  • Shoes gone again!
Re: JRMark -- JRiver's Benchmark Tool
« Reply #300 on: January 05, 2016, 09:54:33 am »

The image creation / destruction is totally the difference.

I've asked our Linux expert to take a look.
Logged
Matt Ashland, JRiver Media Center

bob

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 13578
Re: JRMark -- JRiver's Benchmark Tool
« Reply #301 on: January 05, 2016, 10:16:14 am »

On Linux, latest version 21.0.28:

Code: [Select]
=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 3.316 seconds
    Single-threaded floating point math... 2.285 seconds
    Multi-threaded integer math... 1.680 seconds
    Multi-threaded mixed math... 1.164 seconds
Score: 2250

Running 'Image' benchmark...
    Image creation / destruction... 3.295 seconds
    Flood filling... 0.217 seconds
    Direct copying... 0.355 seconds
    Small renders... 0.732 seconds
    Bilinear rendering... 0.583 seconds
    Bicubic rendering... 0.303 seconds
Score: 4011

Running 'Database' benchmark...
    Create database... 0.228 seconds
    Populate database... 0.901 seconds
    Save database... 0.199 seconds
    Reload database... 0.102 seconds
    Search database... 0.594 seconds
    Sort database... 0.671 seconds
    Group database... 0.432 seconds
Score: 6876

JRMark (version 21.0.28): 4379

I'll switch to windows on this same machine and run it again.

Holy crap that's a big difference:

Code: [Select]
=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 3,321 seconds
    Single-threaded floating point math... 2,284 seconds
    Multi-threaded integer math... 1,933 seconds
    Multi-threaded mixed math... 1,165 seconds
Score: 2183

Running 'Image' benchmark...
    Image creation / destruction... 0,260 seconds
    Flood filling... 0,219 seconds
    Direct copying... 0,339 seconds
    Small renders... 0,809 seconds
    Bilinear rendering... 0,605 seconds
    Bicubic rendering... 0,333 seconds
Score: 8573

Running 'Database' benchmark...
    Create database... 0,139 seconds
    Populate database... 0,863 seconds
    Save database... 0,258 seconds
    Reload database... 0,073 seconds
    Search database... 0,650 seconds
    Sort database... 0,728 seconds
    Group database... 0,491 seconds
Score: 6714

JRMark (version 21.0.30): 5823

32 or 64 bit MC?
Logged

haylo75

  • Regular Member
  • Junior Woodchuck
  • **
  • Posts: 61
Re: JRMark -- JRiver's Benchmark Tool
« Reply #302 on: January 05, 2016, 01:27:19 pm »

My interest is in thumbnail generation, as when I look at server logs I see that thumbs are generated on the fly when using remote clients.  I take it this area would be impacted should I make a move to Linux?
Logged
Kerr

mwillems

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 5181
  • "Linux Merit Badge" Recipient
Re: JRMark -- JRiver's Benchmark Tool
« Reply #303 on: January 05, 2016, 01:31:00 pm »

My interest is in thumbnail generation, as when I look at server logs I see that thumbs are generated on the fly when using remote clients.  I take it this area would be impacted should I make a move to Linux?

You can pre-build the thumbs on a client so that they're generated in advance.  That dramatically improves performance on any client (Windows or Linux),a nd I would recommend it.
Logged

InflatableMouse

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 3978
Re: JRMark -- JRiver's Benchmark Tool
« Reply #304 on: January 05, 2016, 02:19:17 pm »

Logged

haylo75

  • Regular Member
  • Junior Woodchuck
  • **
  • Posts: 61
Re: JRMark -- JRiver's Benchmark Tool
« Reply #305 on: January 05, 2016, 02:21:08 pm »

You can pre-build the thumbs on a client so that they're generated in advance.  That dramatically improves performance on any client (Windows or Linux),a nd I would recommend it.
Sorry, I should've been more specific.  I'm referencing the Gizmo client, (and JRemote too) for which I don't believe this is an option.
Logged
Kerr

eam81969

  • Recent member
  • *
  • Posts: 23
Re: JRMark -- JRiver's Benchmark Tool
« Reply #306 on: January 06, 2016, 08:45:25 pm »

=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 4.634 seconds
    Single-threaded floating point math... 3.187 seconds
    Multi-threaded integer math... 2.601 seconds
    Multi-threaded mixed math... 1.618 seconds
Score: 1578

Running 'Image' benchmark...
    Image creation / destruction... 0.350 seconds
    Flood filling... 0.261 seconds
    Direct copying... 0.409 seconds
    Small renders... 1.082 seconds
    Bilinear rendering... 0.844 seconds
    Bicubic rendering... 0.464 seconds
Score: 6453

Running 'Database' benchmark...
    Create database... 0.189 seconds
    Populate database... 1.070 seconds
    Save database... 0.263 seconds
    Reload database... 0.048 seconds
    Search database... 0.875 seconds
    Sort database... 0.988 seconds
    Group database... 0.715 seconds
Score: 5184

JRMark (version 21.0.23): 4405

To be fair, I just built this machine in the past 2 weeks.

Eric
Logged
i5-6500 6M Skylake Quad-Core 3.2 GHz LGA 1151 w/Intel HD Graphics 530
ASRock H170M Pro4 LGA 1151 Intel H170 HDMI SATA 6Gb/s USB 3.0 Micro ATX Intel Mobo
8GB (2 x 4GB) 288-Pin DDR4 SDRAM DDR4 2133 (PC4 17000)
HGST Deskstar NAS 4TB 7200 RPM 64MB Cache SATA 6.0Gb/s 3.5" HD
DVDFab
Windows 10
JRMark: 4405

bob

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 13578
Re: JRMark -- JRiver's Benchmark Tool
« Reply #307 on: January 07, 2016, 11:09:14 am »

For some reason the 64 bit build is about half the speed of the 32 bit build in the image benchmarks.
Not sure if this translates into a real world difference. We are looking into it.
Logged

InflatableMouse

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 3978
Re: JRMark -- JRiver's Benchmark Tool
« Reply #308 on: January 08, 2016, 12:51:43 am »

For some reason the 64 bit build is about half the speed of the 32 bit build in the image benchmarks.
Not sure if this translates into a real world difference. We are looking into it.


Twice as many bits, half the speed. Makes sense  ;D.

Hope I didn't send you off on another wild goose chase bob!

Thanks for looking into it.
Logged

notbad

  • Recent member
  • *
  • Posts: 8
Re: JRMark -- JRiver's Benchmark Tool
« Reply #309 on: January 13, 2016, 03:14:13 am »

Hello everybody, greeting from Italy...this is my first post!  ;D
I had to upgrade (good excuse yeah!) since my 4 year old ASUS laptop (i7 2660MQ, 8GB ram, JRiver Benchmark at about 1.100) wasn't good enough (very often stopping playback and buffering) to resample 1 bit, 1411,2 kHz files. DAC is a LH LABS GEEK PULSE that doesn't natively support DSD256, but is able to pack it in DOP format).

Here my new build:
asus H170 pro motherboard, intel i5 6600, 16GB ram (2133), samsung 950 pro ssd, WD My Cloud NAS

here the results, but i have a question: does the NAS affect the benchmark test? what if libray was in an internal storage drive?

new build seems not to be very good at "Math"!  :)

=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 4,327 seconds
    Single-threaded floating point math... 3,000 seconds
    Multi-threaded integer math... 2,376 seconds
    Multi-threaded mixed math... 1,479 seconds
Score: 1699

Running 'Image' benchmark...
    Image creation / destruction... 0,411 seconds
    Flood filling... 0,240 seconds
    Direct copying... 0,351 seconds
    Small renders... 1,000 seconds
    Bilinear rendering... 0,776 seconds
    Bicubic rendering... 0,427 seconds
Score: 6864

Running 'Database' benchmark...
    Create database... 0,178 seconds
    Populate database... 0,934 seconds
    Save database... 0,197 seconds
    Reload database... 0,065 seconds
    Search database... 0,830 seconds
    Sort database... 0,911 seconds
    Group database... 0,626 seconds
Score: 5748

JRMark (version 21.0.23): 4770

does the NAS affect test results?

kind regards
Logged

Johnny_Friendly

  • Recent member
  • *
  • Posts: 6
Re: JRMark -- JRiver's Benchmark Tool
« Reply #310 on: January 20, 2016, 10:07:13 pm »

i7-4790k
mobo: Gigabyte G1 Sniper Z87
8GB DDR3 1866 Ram

=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 3.207 seconds
    Single-threaded floating point math... 1.964 seconds
    Multi-threaded integer math... 0.948 seconds
    Multi-threaded mixed math... 0.637 seconds
Score: 2813

Running 'Image' benchmark...
    Image creation / destruction... 0.114 seconds
    Flood filling... 0.261 seconds
    Direct copying... 0.359 seconds
    Small renders... 0.883 seconds
    Bilinear rendering... 0.673 seconds
    Bicubic rendering... 0.362 seconds
Score: 8298

Running 'Database' benchmark...
    Create database... 0.155 seconds
    Populate database... 0.707 seconds
    Save database... 0.096 seconds
    Reload database... 0.029 seconds
    Search database... 0.651 seconds
    Sort database... 0.784 seconds
    Group database... 0.515 seconds
Score: 7321

JRMark (version 21.0.23): 6144
Logged

mwillems

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 5181
  • "Linux Merit Badge" Recipient
Re: JRMark -- JRiver's Benchmark Tool
« Reply #311 on: April 12, 2016, 06:08:37 pm »

Raspberry Pi 3, stock clock speeds, running on an SD card

Code: [Select]
=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 9.217 seconds
    Single-threaded floating point math... 12.566 seconds
    Multi-threaded integer math... 6.066 seconds
    Multi-threaded mixed math... 6.730 seconds
Score: 549

Running 'Image' benchmark...
    Image creation / destruction... 1.965 seconds
    Flood filling... 4.748 seconds
    Direct copying... 4.398 seconds
    Small renders... 2.060 seconds
    Bilinear rendering... 3.675 seconds
    Bicubic rendering... 0.832 seconds
Score: 1245

Running 'Database' benchmark...
    Create database... 4.750 seconds
    Populate database... 9.453 seconds
    Save database... 1.073 seconds
    Reload database... 0.339 seconds
    Search database... 17.877 seconds
    Sort database... 7.876 seconds
    Group database... 4.534 seconds
Score: 468

JRMark (version 21.0.63): 754

The pi's are getting pretty close to some of the lower end intel hardware.  I expect the model 4 (unannounced and hypothetical) will be within striking distance of the Atom/Celerons.
Logged

Matt

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 42034
  • Shoes gone again!
Re: JRMark -- JRiver's Benchmark Tool
« Reply #312 on: April 13, 2016, 08:34:12 am »

A new 6700k system.  I'm running at 4.56 because I'm wimpy.

JRMark (version 21.0.72): 5356

Code: [Select]
=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 3.292 seconds
    Single-threaded floating point math... 2.273 seconds
    Multi-threaded integer math... 0.981 seconds
    Multi-threaded mixed math... 0.619 seconds
Score: 2652

Running 'Image' benchmark...
    Image creation / destruction... 0.357 seconds
    Flood filling... 0.224 seconds
    Direct copying... 0.350 seconds
    Small renders... 0.837 seconds
    Bilinear rendering... 0.664 seconds
    Bicubic rendering... 0.373 seconds
Score: 7843

Running 'Database' benchmark...
    Create database... 0.132 seconds
    Populate database... 0.800 seconds
    Save database... 0.246 seconds
    Reload database... 0.055 seconds
    Search database... 1.269 seconds
    Sort database... 0.842 seconds
    Group database... 0.514 seconds
Score: 5575

JRMark (version 21.0.72): 5356
Logged
Matt Ashland, JRiver Media Center

mojave

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 3732
  • Requires "iTunes or better" so I installed JRiver
Re: JRMark -- JRiver's Benchmark Tool
« Reply #313 on: April 13, 2016, 12:23:55 pm »

I did my build in September of 2014 and scored 6305. I'm curious why it isn't beat by the newer systems? I was only clocked at around 4 Ghz and it wasn't even the fastest CPU available. Maybe it was the M2 SSD drive. Would that make the difference? Maybe JRMark was faster back then.  ;D

Code: [Select]
=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 3.122 seconds
    Single-threaded floating point math... 1.918 seconds
    Multi-threaded integer math... 0.595 seconds
    Multi-threaded mixed math... 0.398 seconds
Score: 3150

Running 'Image' benchmark...
    Image creation / destruction... 0.226 seconds
    Flood filling... 0.258 seconds
    Direct copying... 0.237 seconds
    Small renders... 0.877 seconds
    Bilinear rendering... 0.645 seconds
    Bicubic rendering... 0.349 seconds
Score: 8489

Running 'Database' benchmark...
    Create database... 0.141 seconds
    Populate database... 0.801 seconds
    Save database... 0.210 seconds
    Reload database... 0.039 seconds
    Search database... 0.640 seconds
    Sort database... 0.565 seconds
    Group database... 0.559 seconds
Score: 7275

JRMark (version 20.0.14): 6305
Logged

Matt

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 42034
  • Shoes gone again!
Re: JRMark -- JRiver's Benchmark Tool
« Reply #314 on: April 13, 2016, 02:51:05 pm »

I did my build in September of 2014 and scored 6305. I'm curious why it is beat by the newer systems? I was only clocked at around 4 Ghz and it wasn't even the fastest CPU available. Maybe it was the M2 SSD drive. Would that make the difference. Maybe JRMark was faster back then.  ;D

Well that's sure interesting.  I guess I'll double-check my power saver settings again.  At one point my PC would max out at like 20%.  This was on my old computer, but I have no idea how that ever happened!  I put it back to 100% and everything sure got a lot faster.
Logged
Matt Ashland, JRiver Media Center

mwillems

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 5181
  • "Linux Merit Badge" Recipient
Re: JRMark -- JRiver's Benchmark Tool
« Reply #315 on: April 13, 2016, 07:00:42 pm »

I did my build in September of 2014 and scored 6305. I'm curious why it is beat by the newer systems? I was only clocked at around 4 Ghz and it wasn't even the fastest CPU available. Maybe it was the M2 SSD drive. Would that make the difference. Maybe JRMark was faster back then.  ;D

Code: [Select]
=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 3.122 seconds
    Single-threaded floating point math... 1.918 seconds
    Multi-threaded integer math... 0.595 seconds
    Multi-threaded mixed math... 0.398 seconds
Score: 3150

Running 'Image' benchmark...
    Image creation / destruction... 0.226 seconds
    Flood filling... 0.258 seconds
    Direct copying... 0.237 seconds
    Small renders... 0.877 seconds
    Bilinear rendering... 0.645 seconds
    Bicubic rendering... 0.349 seconds
Score: 8489

Running 'Database' benchmark...
    Create database... 0.141 seconds
    Populate database... 0.801 seconds
    Save database... 0.210 seconds
    Reload database... 0.039 seconds
    Search database... 0.640 seconds
    Sort database... 0.565 seconds
    Group database... 0.559 seconds
Score: 7275

JRMark (version 20.0.14): 6305

Mojave, in the interests of science, could you bench it again?  I have a 4790K that's JRMark is about 800 points lower today than it was last year, and I was kind of suspicious, but didn't know what to think.  I'd like to know if JRMark has normalized downward or if I've really lost that much performance (I changed the overclock, but not that much!).  Independent testing on a system that had previously been tested in v20 would be a help.
Logged

Matt

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 42034
  • Shoes gone again!
Re: JRMark -- JRiver's Benchmark Tool
« Reply #316 on: April 13, 2016, 07:22:22 pm »

Yeah mojave, please bench again.

It's totally possible something screwy happened to JRMark.  We've switched compilers and other internal stuff through the years so it's totally possible there's some unintended consequence.

Having a before and after number would help us know if that's the case.

Thanks.
Logged
Matt Ashland, JRiver Media Center

Matt

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 42034
  • Shoes gone again!
Re: JRMark -- JRiver's Benchmark Tool
« Reply #317 on: April 14, 2016, 05:48:14 am »

I'm comparing MC 19 and MC 21 here.

19.0.165

=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 3.294 seconds
    Single-threaded floating point math... 2.279 seconds
    Multi-threaded integer math... 1.047 seconds
    Multi-threaded mixed math... 0.696 seconds
Score: 2597

Running 'Image' benchmark...
    Image creation / destruction... 0.419 seconds
    Flood filling... 0.229 seconds
    Direct copying... 0.377 seconds
    Small renders... 0.844 seconds
    Bilinear rendering... 0.742 seconds
    Bicubic rendering... 0.419 seconds
Score: 7262

Running 'Database' benchmark...
    Create database... 0.141 seconds
    Populate database... 0.859 seconds
    Save database... 0.254 seconds
    Reload database... 0.048 seconds
    Search database... 0.652 seconds
    Sort database... 0.723 seconds
    Group database... 0.471 seconds
Score: 6831

JRMark (version 19.0.165): 5563

And using the latest MC 21.  It's close, but not as close as I would have expected:

=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 3.292 seconds
    Single-threaded floating point math... 2.274 seconds
    Multi-threaded integer math... 0.991 seconds
    Multi-threaded mixed math... 0.761 seconds
Score: 2596

Running 'Image' benchmark...
    Image creation / destruction... 0.338 seconds
    Flood filling... 0.224 seconds
    Direct copying... 0.367 seconds
    Small renders... 0.843 seconds
    Bilinear rendering... 0.742 seconds
    Bicubic rendering... 0.435 seconds
Score: 7461

Running 'Database' benchmark...
    Create database... 0.145 seconds
    Populate database... 0.821 seconds
    Save database... 0.237 seconds
    Reload database... 0.049 seconds
    Search database... 1.270 seconds
    Sort database... 0.855 seconds
    Group database... 0.502 seconds
Score: 5542

JRMark (version 21.0.72): 5200

The "Search database..." entry went from 0.652 seconds to 1.270 seconds.

Interesting...
Logged
Matt Ashland, JRiver Media Center

Matt

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 42034
  • Shoes gone again!
Re: JRMark -- JRiver's Benchmark Tool
« Reply #318 on: April 14, 2016, 06:29:12 am »

I've got a fix for the search speed change pending.  I'm asking the team for feedback before I check in.
Logged
Matt Ashland, JRiver Media Center

mojave

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 3732
  • Requires "iTunes or better" so I installed JRiver
Re: JRMark -- JRiver's Benchmark Tool
« Reply #319 on: April 14, 2016, 08:50:41 am »

I'll bench again . . . after I finish my taxes.  :)
Logged

Matt

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 42034
  • Shoes gone again!
Re: JRMark -- JRiver's Benchmark Tool
« Reply #320 on: April 14, 2016, 09:08:58 am »

The same 6700k system, but with a fixed Media Center.

JRMark (version 21.0.73): 5565

Code: [Select]
=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 3.293 seconds
    Single-threaded floating point math... 2.273 seconds
    Multi-threaded integer math... 1.094 seconds
    Multi-threaded mixed math... 0.749 seconds
Score: 2564

Running 'Image' benchmark...
    Image creation / destruction... 0.310 seconds
    Flood filling... 0.228 seconds
    Direct copying... 0.371 seconds
    Small renders... 0.835 seconds
    Bilinear rendering... 0.727 seconds
    Bicubic rendering... 0.423 seconds
Score: 7597

Running 'Database' benchmark...
    Create database... 0.134 seconds
    Populate database... 0.820 seconds
    Save database... 0.219 seconds
    Reload database... 0.045 seconds
    Search database... 0.675 seconds
    Sort database... 0.857 seconds
    Group database... 0.541 seconds
Score: 6532

JRMark (version 21.0.73): 5565
Logged
Matt Ashland, JRiver Media Center

)p(

  • Citizen of the Universe
  • *****
  • Posts: 579
Re: JRMark -- JRiver's Benchmark Tool
« Reply #321 on: April 28, 2016, 03:52:59 am »

So lately I run an osx virtual machine on my windows pc and I was interested to see how their jriver benchmark scores compared. Its an i7 - 4770 @ 3.4 ghz with 36gb ddr3 ram at 1600.

The virtual mac with the latest jriver build:

=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
Single-threaded integer math... 3.789 seconds
Single-threaded floating point math... 2.358 seconds
Multi-threaded integer math... 1.027 seconds
Multi-threaded mixed math... 0.650 seconds
Score: 2429

Running 'Image' benchmark...
Image creation / destruction... 1.333 seconds
Flood filling... 0.389 seconds
Direct copying... 2.704 seconds
Small renders... 0.942 seconds
Bilinear rendering... 1.213 seconds
Bicubic rendering... 0.454 seconds
Score: 3127

Running 'Database' benchmark...
Create database... 0.245 seconds
Populate database... 1.033 seconds
Save database... 0.137 seconds
Reload database... 0.057 seconds
Search database... 1.577 seconds
Sort database... 0.737 seconds
Group database... 0.677 seconds
Score: 4816

JRMark (version 21.0.66): 3457


And with Windows 10:

=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 3,768 seconds
    Single-threaded floating point math... 2,319 seconds
    Multi-threaded integer math... 1,242 seconds
    Multi-threaded mixed math... 0,728 seconds
Score: 2358

Running 'Image' benchmark...
    Image creation / destruction... 0,357 seconds
    Flood filling... 0,313 seconds
    Direct copying... 0,424 seconds
    Small renders... 1,068 seconds
    Bilinear rendering... 0,823 seconds
    Bicubic rendering... 0,461 seconds
Score: 6385

Running 'Database' benchmark...
    Create database... 0,172 seconds
    Populate database... 0,877 seconds
    Save database... 0,145 seconds
    Reload database... 0,058 seconds
    Search database... 1,642 seconds
    Sort database... 0,858 seconds
    Group database... 0,570 seconds
Score: 4974

JRMark (version 21.0.72): 4572


The math and database scores are pretty much identical between a few benchmark runs.
But in the graphics test the virtual machine gets only about 50% of the windows 10 host score.
The difference is in the image creation/destruction and in the Direct copying tests.
Logged

pschelbert

  • Galactic Citizen
  • ****
  • Posts: 459
Re: JRMark -- JRiver's Benchmark Tool
« Reply #322 on: June 12, 2016, 04:46:40 pm »

Hi

old machine:
JRMark (version 21.0.80): 3103
I7P860, 2.8GHz, 8GB RAM, Win7Pro64
Asus MB P7P55D-E

HTPC (fanless) new
JRMark (version 21.0.80): 4826
I76700T, 2.8GHz, 16MB RAM,
Asus MB H170I Plus D3

looks to be good.

Peter

Logged

pschelbert

  • Galactic Citizen
  • ****
  • Posts: 459
Re: JRMark -- JRiver's Benchmark Tool
« Reply #323 on: June 13, 2016, 07:03:26 am »

Hi

i went depper into it. Surprisingly the old processor I7P860 has in some parts even more power than the new I76700T.
No progress what processing power is concerned.
TDPD is 1/3 however of the new processor, so possible to run fanless, thats an advantage.

old machine:
JRMark (version 21.0.80): 3103
I7P860, 2.8GHz, 8GB RAM, Win7Pro64
Asus MB P7P55D-E

Running 'Math' benchmark...
    Single-threaded integer math... 4.512 seconds
    Single-threaded floating point math... 2.490 seconds
    Multi-threaded integer math... 1.366 seconds
    Multi-threaded mixed math... 0.914 seconds
Score: 2047


new machine:
HTPC (fanless) new
JRMark (version 21.0.80): 4826
I76700T, 2.8GHz, 16MB RAM, Win7Pro64
Asus MB H170I Plus D3

Running 'Math' benchmark...
    Single-threaded integer math... 4.342 seconds
    Single-threaded floating point math... 2.969 seconds
    Multi-threaded integer math... 1.304 seconds
    Multi-threaded mixed math... 0.777 seconds
Score: 2023


Peter
Logged

Hilton

  • Regular Member
  • Citizen of the Universe
  • *****
  • Posts: 1291
Re: JRMark -- JRiver's Benchmark Tool
« Reply #324 on: July 08, 2016, 08:59:27 am »

Got a new surface book to take on holidays. Wow what amazing battery life! Pretty versatile little machine.


i7 16GB 512GB SSD

=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 4.864 seconds
    Single-threaded floating point math... 3.397 seconds
    Multi-threaded integer math... 2.859 seconds
    Multi-threaded mixed math... 1.970 seconds
Score: 1452

Running 'Image' benchmark...
    Image creation / destruction... 0.406 seconds
    Flood filling... 0.280 seconds
    Direct copying... 0.487 seconds
    Small renders... 1.266 seconds
    Bilinear rendering... 1.543 seconds
    Bicubic rendering... 0.977 seconds
Score: 4436

Running 'Database' benchmark...
    Create database... 0.213 seconds
    Populate database... 1.139 seconds
    Save database... 0.323 seconds
    Reload database... 0.073 seconds
    Search database... 0.981 seconds
    Sort database... 0.964 seconds
    Group database... 0.659 seconds
Score: 4940

JRMark (version 21.0.90): 3609
Logged

mwillems

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 5181
  • "Linux Merit Badge" Recipient
Re: JRMark -- JRiver's Benchmark Tool
« Reply #325 on: July 08, 2016, 09:20:29 am »

How're you liking the Surface Book so far Hilton?  How's the "clipboard" mode?  I almost grabbed one, but got the Surface Pro 4 instead and I'm curious how the "other half" live.

Looks like the JRMark performance is pretty similar, so the main differentiator seems to be the form factor and the Book's much improved battery life.

For comparison, here's the bench on the i7 Surface Pro 4 with 16GB RAM:

Quote
=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 4.704 seconds
    Single-threaded floating point math... 3.118 seconds
    Multi-threaded integer math... 2.686 seconds
    Multi-threaded mixed math... 1.829 seconds
Score: 1540

Running 'Image' benchmark...
    Image creation / destruction... 0.362 seconds
    Flood filling... 0.187 seconds
    Direct copying... 0.426 seconds
    Small renders... 1.143 seconds
    Bilinear rendering... 1.496 seconds
    Bicubic rendering... 0.923 seconds
Score: 4848

Running 'Database' benchmark...
    Create database... 0.192 seconds
    Populate database... 1.171 seconds
    Save database... 0.370 seconds
    Reload database... 0.088 seconds
    Search database... 0.947 seconds
    Sort database... 1.083 seconds
    Group database... 0.766 seconds
Score: 4657

JRMark (version 21.0.90): 3681
Logged

Hilton

  • Regular Member
  • Citizen of the Universe
  • *****
  • Posts: 1291
Re: JRMark -- JRiver's Benchmark Tool
« Reply #326 on: July 08, 2016, 09:43:59 am »

Early days but I do like it better on my lap and much easier to type. The track pad is great, the extra larger screen (21% bigger than Pro 4) I also noticed I sit back in my chair more rather than leaning forward.

More comfortable form factor I think. (coming from a Surface Pro 2)

Had to install newer NVME M.2 express drivers to get the SSD singing.. was abysmally slow with 80mb/s write and only 500mb/s read with factory drivers. Now almost 600MB/s write and 1400MB/s read and it feels much snappier!

Can do some light gaming pretty well too with graphics on high and down scaled to 1600x900.

It's a really nice machine.  The Surface Pro 2 was too small and heavy to be an effective tablet, the 4 looks like the perfect form factor for tablet and quite workable as laptop but there just hasn't been enough apps I've wanted to use in tablet mode to justify a tablet/laptop so I've gone the other way with laptop/tablet with better GPU performance and battery life.  So far soo good! Will be good to store/watch movies on too for holiday in FiJi in 2 weeks!

I now prefer the Android as a pure tablet and have both a 7" and 8" and think the 8" is perfect and not so heavy that I cant just throw it in the bag with the surface.
The iPhone 6s+ fits the other end of the scale for me.

 PS. Forgot to answer your clipboard question. ....... Downside, there's no stand, if you want to use it in a free standing position without keyboard your out of luck.... but it is reversible on the keyboard.  It's still reasonably light for it's size (about 700gms) but almost too big and uncomfortable to use in handheld tablet mode for any length of time.  It works really well in reversed mode folded over the keyboard for just general web browsing on your lap though.
Logged

WolfAudioSystems

  • Regular Member
  • Recent member
  • *
  • Posts: 23
  • It can always sound better
Re: JRMark -- JRiver's Benchmark Tool
« Reply #327 on: July 11, 2016, 01:29:57 pm »

=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 4.425 seconds
    Single-threaded floating point math... 3.069 seconds
    Multi-threaded integer math... 2.380 seconds
    Multi-threaded mixed math... 1.661 seconds
Score: 1647

Running 'Image' benchmark...
    Image creation / destruction... 0.383 seconds
    Flood filling... 0.234 seconds
    Direct copying... 0.357 seconds
    Small renders... 1.098 seconds
    Bilinear rendering... 0.856 seconds
    Bicubic rendering... 0.480 seconds
Score: 6456

Running 'Database' benchmark...
    Create database... 0.176 seconds
    Populate database... 1.119 seconds
    Save database... 0.356 seconds
    Reload database... 0.102 seconds
    Search database... 0.876 seconds
    Sort database... 1.113 seconds
    Group database... 0.691 seconds
Score: 4850

JRMark (version 21.0.83): 4318
Logged
Wolf Audio Systems - Wolf Alpha II - Audiophile HTPC
Arcam AVR450 - Ayre Codex - Kef R series - Tributaries Series 8 Cabling

Hilton

  • Regular Member
  • Citizen of the Universe
  • *****
  • Posts: 1291
Re: JRMark -- JRiver's Benchmark Tool
« Reply #328 on: July 17, 2016, 06:51:33 am »

Fresh build of windows 10 64bit.

Intel i7 4960X @ 4.8Ghz
Ram @ 2400Mhz
GTX 1080 @ 2150Mhz
2 x 120GB Sandisk Extreme SSD in RAID 0

This runs MadVR @ 64/32 NNEDI3 1080p and 128/32 NNEDI3 for DVD
Real nice picture improvement over the GTX 580. :)


=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 2.819 seconds
    Single-threaded floating point math... 1.880 seconds
    Multi-threaded integer math... 0.594 seconds
    Multi-threaded mixed math... 0.398 seconds
Score: 3339

Running 'Image' benchmark...
    Image creation / destruction... 0.340 seconds
    Flood filling... 0.367 seconds
    Direct copying... 0.244 seconds
    Small renders... 0.951 seconds
    Bilinear rendering... 0.689 seconds
    Bicubic rendering... 0.438 seconds
Score: 7263

Running 'Database' benchmark...
    Create database... 0.149 seconds
    Populate database... 0.846 seconds
    Save database... 0.195 seconds
    Reload database... 0.051 seconds
    Search database... 0.722 seconds
    Sort database... 0.776 seconds
    Group database... 0.508 seconds
Score: 6621

JRMark (version 21.0.90): 5741
Logged

Hilton

  • Regular Member
  • Citizen of the Universe
  • *****
  • Posts: 1291
Re: JRMark -- JRiver's Benchmark Tool
« Reply #329 on: July 17, 2016, 06:54:49 am »

just realised that's with speed step dropping the CPU back.. will have to redo with fixed clock.

That's better!

=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 2.819 seconds
    Single-threaded floating point math... 1.880 seconds
    Multi-threaded integer math... 0.594 seconds
    Multi-threaded mixed math... 0.398 seconds
Score: 3339

Running 'Image' benchmark...
    Image creation / destruction... 0.340 seconds
    Flood filling... 0.367 seconds
    Direct copying... 0.244 seconds
    Small renders... 0.951 seconds
    Bilinear rendering... 0.689 seconds
    Bicubic rendering... 0.438 seconds
Score: 7263

Running 'Database' benchmark...
    Create database... 0.149 seconds
    Populate database... 0.846 seconds
    Save database... 0.195 seconds
    Reload database... 0.051 seconds
    Search database... 0.722 seconds
    Sort database... 0.776 seconds
    Group database... 0.508 seconds
Score: 6621

JRMark (version 21.0.90): 5741
Logged

kr4

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 723
Re: JRMark -- JRiver's Benchmark Tool
« Reply #330 on: July 19, 2016, 01:11:51 pm »

My i7-4770 running at 3.10 on MC22:
Running 'Math' benchmark...
    Single-threaded integer math... 3.953 seconds
    Single-threaded floating point math... 2.430 seconds
    Multi-threaded integer math... 1.190 seconds
    Multi-threaded mixed math... 0.852 seconds
Score: 2255

Running 'Image' benchmark...
    Image creation / destruction... 0.146 seconds
    Flood filling... 0.323 seconds
    Direct copying... 0.445 seconds
    Small renders... 1.099 seconds
    Bilinear rendering... 1.093 seconds
    Bicubic rendering... 0.570 seconds
Score: 5984

Running 'Database' benchmark...
    Create database... 0.172 seconds
    Populate database... 1.039 seconds
    Save database... 0.296 seconds
    Reload database... 0.096 seconds
    Search database... 0.869 seconds
    Sort database... 0.874 seconds
    Group database... 0.523 seconds
Score: 5558

JRMark (version 22.0.9): 4599

Time to upgrade....................................
Logged
Kal Rubinson
"Music in the Round"
Senior Contributing Editor, Stereophile

JimH

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 71619
  • Where did I put my teeth?
Re: JRMark -- JRiver's Benchmark Tool
« Reply #331 on: July 19, 2016, 02:31:59 pm »

That's a very good machine. 
Logged

Hilton

  • Regular Member
  • Citizen of the Universe
  • *****
  • Posts: 1291
Re: JRMark -- JRiver's Benchmark Tool
« Reply #332 on: August 05, 2016, 11:39:01 pm »

With the new GTX1080 SLI and Intel 750 NVMe 1.2TB SSD. Not much of an improvement but still faster.

=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 2.820 seconds
    Single-threaded floating point math... 1.877 seconds
    Multi-threaded integer math... 0.716 seconds
    Multi-threaded mixed math... 0.430 seconds
Score: 3252

Running 'Image' benchmark...
    Image creation / destruction... 0.244 seconds
    Flood filling... 0.369 seconds
    Direct copying... 0.244 seconds
    Small renders... 0.952 seconds
    Bilinear rendering... 0.656 seconds
    Bicubic rendering... 0.396 seconds
Score: 7686

Running 'Database' benchmark...
    Create database... 0.146 seconds
    Populate database... 0.813 seconds
    Save database... 0.191 seconds
    Reload database... 0.034 seconds
    Search database... 0.727 seconds
    Sort database... 0.788 seconds
    Group database... 0.468 seconds
Score: 6788

JRMark (version 21.0.90): 5909
Logged

bhampster

  • World Citizen
  • ***
  • Posts: 180
Re: JRMark -- JRiver's Benchmark Tool
« Reply #333 on: August 16, 2016, 06:09:33 am »

Apparently 2 GPU's are better than 1.

I have a i5 with integrated HD4600 graphics which are usually disabled because I added a GTX750.

I enabled the integrated graphics in the BIOS and my JMark score changed from 4087 to 4095.

Wanted to let people know enabling integrated GPU may help JMark scores... maybe even resulting in better performance. :)

I surely love no cost upgrades ! It's very cool that JRiver is so optimized these DX12 enhancements result in better JRiver performance as well.
 ;D
Logged

Nikkel

  • Recent member
  • *
  • Posts: 35
Re: JRMark -- JRiver's Benchmark Tool
« Reply #334 on: September 15, 2016, 01:41:28 pm »

i5 6600k @4.2GHz, Z170-K, R9 280X, Crucial MX100 SSD

Running 'Math' benchmark...
    Single-threaded integer math... 3,622 seconds
    Single-threaded floating point math... 2,489 seconds
    Multi-threaded integer math... 1,822 seconds
    Multi-threaded mixed math... 1,249 seconds
Score: 2069

Running 'Image' benchmark...
    Image creation / destruction... 0,319 seconds
    Flood filling... 0,218 seconds
    Direct copying... 0,340 seconds
    Small renders... 0,893 seconds
    Bilinear rendering... 0,676 seconds
    Bicubic rendering... 0,382 seconds
Score: 7784

Running 'Database' benchmark...
    Create database... 0,143 seconds
    Populate database... 0,906 seconds
    Save database... 0,304 seconds
    Reload database... 0,130 seconds
    Search database... 0,733 seconds
    Sort database... 0,696 seconds
    Group database... 0,496 seconds
Score: 6308

JRMark (version 22.0.21): 5387
Logged

jmone

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 14277
  • I won! I won!
Re: JRMark -- JRiver's Benchmark Tool
« Reply #335 on: September 17, 2016, 10:10:16 pm »

It was over 4 years ago the 5,000+ JR Mark was broken, yet it seems that there is relatively little improvement.  Some 6K+ benchmarks, but ... I'd have thought the improvement over this time would have been multiples.
Logged
JRiver CEO Elect

RoderickGI

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 8186
Re: JRMark -- JRiver's Benchmark Tool
« Reply #336 on: September 18, 2016, 12:05:01 am »

It was over 4 years ago the 5,000+ JR Mark was broken, yet it seems that there is relatively little improvement.  Some 6K+ benchmarks, but ... I'd have thought the improvement over this time would have been multiples.

This is probably the reason. I've added a discrete graphics card since last reporting my Benchmark, and I now have a significantly lower score.

Now:
=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 3.686 seconds
    Single-threaded floating point math... 2.450 seconds
    Multi-threaded integer math... 1.911 seconds
    Multi-threaded mixed math... 1.348 seconds
Score: 2022

Running 'Image' benchmark...
    Image creation / destruction... 1.345 seconds
    Flood filling... 0.395 seconds
    Direct copying... 0.522 seconds
    Small renders... 1.243 seconds
    Bilinear rendering... 0.891 seconds
    Bicubic rendering... 0.535 seconds
Score: 4462

Running 'Database' benchmark...
    Create database... 0.243 seconds
    Populate database... 0.977 seconds
    Save database... 0.131 seconds
    Reload database... 0.081 seconds
    Search database... 1.044 seconds
    Sort database... 1.306 seconds
    Group database... 0.736 seconds
Score: 4760

JRMark (version 22.0.25): 3748

Then: http://yabb.jriver.com/interact/index.php/topic,54396.msg621634/topicseen.html#msg621634
=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 3.701 seconds
    Single-threaded floating point math... 2.453 seconds
    Multi-threaded integer math... 1.903 seconds
    Multi-threaded mixed math... 1.270 seconds
Score: 2037

Running 'Image' benchmark...
    Image creation / destruction... 0.169 seconds
    Flood filling... 0.391 seconds
    Direct copying... 0.472 seconds
    Small renders... 1.215 seconds
    Bilinear rendering... 0.863 seconds
    Bicubic rendering... 0.497 seconds
Score: 6099

Running 'Database' benchmark...
    Create database... 0.261 seconds
    Populate database... 1.023 seconds
    Save database... 0.129 seconds
    Reload database... 0.035 seconds
    Search database... 0.934 seconds
    Sort database... 0.918 seconds
    Group database... 0.588 seconds
Score: 5528

JRMark (version 19.0.156): 4555
Logged
What specific version of MC you are running:MC27.0.27 @ Oct 27, 2020 and updating regularly Jim!                        MC Release Notes: https://wiki.jriver.com/index.php/Release_Notes
What OS(s) and Version you are running:     Windows 10 Pro 64bit Version 2004 (OS Build 19041.572).
The JRMark score of the PC with an issue:    JRMark (version 26.0.52 64 bit): 3419
Important relevant info about your environment:     
  Using the HTPC as a MC Server & a Workstation as a MC Client plus some DLNA clients.
  Running JRiver for Android, JRemote2, Gizmo, & MO 4Media on a Sony Xperia XZ Premium Android 9.
  Playing video out to a Sony 65" TV connected via HDMI, playing digital audio out via motherboard sound card, PCIe TV tuner

Daydream

  • Citizen of the Universe
  • *****
  • Posts: 770
Re: JRMark -- JRiver's Benchmark Tool
« Reply #337 on: September 20, 2016, 08:23:43 pm »

Running 'Image' benchmark...
    Image creation / destruction... 1.345 seconds

That looks suspiciously bad. If all other things remain the same, adding a discrete GPU made those numbers 8 times worse? Hmmm...

That being said, this benchmark used to be a fun exercise. But now the numbers are not comparable anymore. One would need a scorched-Earth config to beat numbers from MC19-20 era.
The question is, should we investigate, or the cause is known?
Logged

RoderickGI

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 8186
Re: JRMark -- JRiver's Benchmark Tool
« Reply #338 on: September 20, 2016, 08:47:34 pm »

That looks suspiciously bad. If all other things remain the same, adding a discrete GPU made those numbers 8 times worse? Hmmm...

You are right. I overlooked that one bad number with my quick visual comparison.

Maybe I should look into that number...driver update, turn off Intel iGPU in BIOS, something else maybe.
Logged
What specific version of MC you are running:MC27.0.27 @ Oct 27, 2020 and updating regularly Jim!                        MC Release Notes: https://wiki.jriver.com/index.php/Release_Notes
What OS(s) and Version you are running:     Windows 10 Pro 64bit Version 2004 (OS Build 19041.572).
The JRMark score of the PC with an issue:    JRMark (version 26.0.52 64 bit): 3419
Important relevant info about your environment:     
  Using the HTPC as a MC Server & a Workstation as a MC Client plus some DLNA clients.
  Running JRiver for Android, JRemote2, Gizmo, & MO 4Media on a Sony Xperia XZ Premium Android 9.
  Playing video out to a Sony 65" TV connected via HDMI, playing digital audio out via motherboard sound card, PCIe TV tuner

jmone

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 14277
  • I won! I won!
Re: JRMark -- JRiver's Benchmark Tool
« Reply #339 on: September 23, 2016, 01:38:18 am »

Seems about the same to me over the 4 years.  Here is what I get now (OC'ed i7-3770k) and what I was getting 4 years ago..... Still amazed the newer PC's are not doing much much better than this however.

Code: [Select]
=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 2.968 seconds
    Single-threaded floating point math... 1.961 seconds
    Multi-threaded integer math... 1.001 seconds
    Multi-threaded mixed math... 0.642 seconds
Score: 2891

Running 'Image' benchmark...
    Image creation / destruction... 0.312 seconds
    Flood filling... 0.296 seconds
    Direct copying... 0.327 seconds
    Small renders... 0.995 seconds
    Bilinear rendering... 0.744 seconds
    Bicubic rendering... 0.430 seconds
Score: 7086

Running 'Database' benchmark...
    Create database... 0.197 seconds
    Populate database... 0.770 seconds
    Save database... 0.108 seconds
    Reload database... 0.066 seconds
    Search database... 0.758 seconds
    Sort database... 1.034 seconds
    Group database... 0.611 seconds
Score: 6068

JRMark (version 22.0.29): 5348

No probs breaking 5,000 with V98 and Asus AutoTune (no manual tweaks by me).
 
=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 3.032 seconds
    Single-threaded floating point math... 2.003 seconds
    Multi-threaded integer math... 0.922 seconds
    Multi-threaded mixed math... 0.599 seconds
Score: 2898

Running 'Image' benchmark...
    Image creation / destruction... 0.203 seconds
    Flood filling... 0.342 seconds
    Direct copying... 0.514 seconds
    Small renders... 0.849 seconds
    Bilinear rendering... 0.662 seconds
    Bicubic rendering... 0.582 seconds
Score: 6979

Running 'Database' benchmark...
    Create database... 0.315 seconds
    Populate database... 1.103 seconds
    Save database... 0.237 seconds
    Reload database... 0.092 seconds
    Search database... 0.755 seconds
    Sort database... 0.776 seconds
    Group database... 0.451 seconds
Score: 5768

JRMark (version 18.0.98): 5215
Logged
JRiver CEO Elect

Daydream

  • Citizen of the Universe
  • *****
  • Posts: 770
Re: JRMark -- JRiver's Benchmark Tool
« Reply #340 on: September 24, 2016, 08:40:06 pm »

Like I said, me, I'm seeing a revers of fortunes. Same hardware (i7 6700K OCed to 4400Mhz, 16GB DDR4 3400, GTX 970, SSDs, the works).

Code: [Select]
=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
Single-threaded integer math... 3.460 seconds
Single-threaded floating point math... 2.390 seconds
Multi-threaded integer math... 1.194 seconds
Multi-threaded mixed math... 0.818 seconds
Score: 2417

Running 'Image' benchmark...
Image creation / destruction... 0.343 seconds
Flood filling... 0.220 seconds
Direct copying... 0.350 seconds
Small renders... 0.869 seconds
Bilinear rendering... 0.703 seconds
Bicubic rendering... 0.435 seconds
Score: 7533

Running 'Database' benchmark...
Create database... 0.141 seconds
Populate database... 0.812 seconds
Save database... 0.198 seconds
Reload database... 0.064 seconds
Search database... 0.671 seconds
Sort database... 0.738 seconds
Group database... 0.596 seconds
Score: 6676

JRMark (version 22.0.28): 5542




Code: [Select]
=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
Single-threaded integer math... 3.473 seconds
Single-threaded floating point math... 2.388 seconds
Multi-threaded integer math... 1.159 seconds
Multi-threaded mixed math... 0.770 seconds
Score: 2439

Running 'Image' benchmark...
Image creation / destruction... 0.277 seconds
Flood filling... 0.230 seconds
Direct copying... 0.349 seconds
Small renders... 0.865 seconds
Bilinear rendering... 0.721 seconds
Bicubic rendering... 0.402 seconds
Score: 7736

Running 'Database' benchmark...
Create database... 0.151 seconds
Populate database... 0.857 seconds
Save database... 0.211 seconds
Reload database... 0.031 seconds
Search database... 0.645 seconds
Sort database... 0.572 seconds
Group database... 0.469 seconds
Score: 7324

JRMark (version 19.0.163): 5833

So with the huge emphasis ON MY SYSTEM (so take it as such), that's about %5 worse overall for version 22. While the Math benchmark holds steady more or less, the Image benchmark is starting to drift away (Image creation/destruction takes 123% time compared to older version), and database benchmark is, ahem, way off (reload database 206% more time, sort and group database 129% and 127% respectively).

This may mean something or nothing depending on a kazilion of factors, that benchmarks on one system can't account for. And all these is for pure statistical fun, I can't say I notice a difference in real time usage.



Second thing, I'd like to know how on God's green Earth did mojave scored 6305 with a Haswell-e i7-5820K (on MC 20)? If that results holds solid, we can draw the conclusion that MC is very well threaded, so the more cores the merrier, speed be damned (the hexacore Haswell-e i7-5820K is 3.3-3.6Ghz, Skylake i7-6700K is 4-4.2Ghz).
Logged

st1

  • Recent member
  • *
  • Posts: 30
Re: JRMark -- JRiver's Benchmark Tool
« Reply #341 on: October 03, 2016, 12:02:48 pm »

intel NUC i7RYH  ubuntu 14.04 64bit  &  JRiver 21
math: 1869
image: 1744
db: 5862
JRMark (version 21.0.4):  3158

UPDATE ubuntu 16.04 64bit JRiver 21
Running 'Math' benchmark...   Score: 1759
Running 'Image' benchmark... Score: 3813
Running 'Database' benchmark.. Score: 4646
JRMark (version 21.0.88): 3406

UPDATE  ubuntu 16.04.1   64 bit  STUDIO
Running 'Math' benchmark...   Score: 1759   to 1874
Running 'Image' benchmark... Score: 3813 to 5123
Running 'Database' benchmark.. Score: 4646 to 5472
JRMark (version 22.0.30): 3406 to 4156

UPDATE  08/2018  ubuntu 18.04.1   64 bit  STUDIO & latest MC24
Running 'Math' benchmark... Score: 1862
Running 'Image' benchmark... Score: 3396
Running 'Database' benchmark... Score: 5504
JRMark (version 24.0.46 64 bit): 3587

-> image performance dropped, but why ?
Logged

bob

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 13578
Re: JRMark -- JRiver's Benchmark Tool
« Reply #342 on: October 03, 2016, 12:30:36 pm »

Like I said, me, I'm seeing a revers of fortunes. Same hardware (i7 6700K OCed to 4400Mhz, 16GB DDR4 3400, GTX 970, SSDs, the works).

Code: [Select]
=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
Single-threaded integer math... 3.460 seconds
Single-threaded floating point math... 2.390 seconds
Multi-threaded integer math... 1.194 seconds
Multi-threaded mixed math... 0.818 seconds
Score: 2417

Running 'Image' benchmark...
Image creation / destruction... 0.343 seconds
Flood filling... 0.220 seconds
Direct copying... 0.350 seconds
Small renders... 0.869 seconds
Bilinear rendering... 0.703 seconds
Bicubic rendering... 0.435 seconds
Score: 7533

Running 'Database' benchmark...
Create database... 0.141 seconds
Populate database... 0.812 seconds
Save database... 0.198 seconds
Reload database... 0.064 seconds
Search database... 0.671 seconds
Sort database... 0.738 seconds
Group database... 0.596 seconds
Score: 6676

JRMark (version 22.0.28): 5542




Code: [Select]
=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
Single-threaded integer math... 3.473 seconds
Single-threaded floating point math... 2.388 seconds
Multi-threaded integer math... 1.159 seconds
Multi-threaded mixed math... 0.770 seconds
Score: 2439

Running 'Image' benchmark...
Image creation / destruction... 0.277 seconds
Flood filling... 0.230 seconds
Direct copying... 0.349 seconds
Small renders... 0.865 seconds
Bilinear rendering... 0.721 seconds
Bicubic rendering... 0.402 seconds
Score: 7736

Running 'Database' benchmark...
Create database... 0.151 seconds
Populate database... 0.857 seconds
Save database... 0.211 seconds
Reload database... 0.031 seconds
Search database... 0.645 seconds
Sort database... 0.572 seconds
Group database... 0.469 seconds
Score: 7324

JRMark (version 19.0.163): 5833

So with the huge emphasis ON MY SYSTEM (so take it as such), that's about %5 worse overall for version 22. While the Math benchmark holds steady more or less, the Image benchmark is starting to drift away (Image creation/destruction takes 123% time compared to older version), and database benchmark is, ahem, way off (reload database 206% more time, sort and group database 129% and 127% respectively).

This may mean something or nothing depending on a kazilion of factors, that benchmarks on one system can't account for. And all these is for pure statistical fun, I can't say I notice a difference in real time usage.



Second thing, I'd like to know how on God's green Earth did mojave scored 6305 with a Haswell-e i7-5820K (on MC 20)? If that results holds solid, we can draw the conclusion that MC is very well threaded, so the more cores the merrier, speed be damned (the hexacore Haswell-e i7-5820K is 3.3-3.6Ghz, Skylake i7-6700K is 4-4.2Ghz).

Windows I assume. I wonder if the dev studio 2015 compile is responsible?

In linux I get the essentially the same results running side by side with MC 22 just a bit faster.

Logged

Hendrik

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 10776
Re: JRMark -- JRiver's Benchmark Tool
« Reply #343 on: October 03, 2016, 12:46:45 pm »

Windows I assume. I wonder if the dev studio 2015 compile is responsible?

There is only a few outliers in there, that is mostly: Image Creation, and Library Sort/Group.
Library sorting and grouping were changed a while ago for a few more features, which takes extra time - especially if you compare 19 to 22, like here.
Logged
~ nevcairiel
~ Author of LAV Filters

bob

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 13578
Re: JRMark -- JRiver's Benchmark Tool
« Reply #344 on: October 03, 2016, 01:23:17 pm »

There is only a few outliers in there, that is mostly: Image Creation, and Library Sort/Group.
Library sorting and grouping were changed a while ago for a few more features, which takes extra time - especially if you compare 19 to 22, like here.
The image creation made me suspicious since there seems like there might have been a change in there somewhere that's affecting the spectrum analyzer in linux. Users reported it worked faster a couple of months ago, still trying to track that one down...
Logged

Manfred

  • Citizen of the Universe
  • *****
  • Posts: 1027
Re: JRMark -- JRiver's Benchmark Tool
« Reply #345 on: November 18, 2016, 08:09:34 am »

I replaced my 10 year old core2 6600 with my HTPC MB, CPU, RAM and (I thought got something smaller an i3 ?):


MC 22.0.41 JRMark Summary:

4435  HTPC:  i3-6100; Asus B150I Pro Gaming; 8GB CL14 2133 MHz RAM; intel m.2 ssd; EVGA 960 SSC
4207  PC:      i7-4770T; MSI Z87M Gaming; 8 GB CL9 1600 MHz RAM; Samsung Evo SSD; MSI 1070 Gaming X: 
1613: OldPC: Core2 6600, Nvidia nForce, 6 GB RAM 800 MHz:

(No details for the Core2 6600)

i3-6100; Asus B150I Pro Gaming; 8GB CL 2133 MHz RAM: 4435

=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 4,135 seconds
    Single-threaded floating point math... 2,865 seconds
    Multi-threaded integer math... 2,196 seconds
    Multi-threaded mixed math... 1,587 seconds
Score: 1762

Running 'Image' benchmark...
    Image creation / destruction... 0,335 seconds
    Flood filling... 0,239 seconds
    Direct copying... 0,402 seconds
    Small renders... 1,060 seconds
    Bilinear rendering... 1,286 seconds
    Bicubic rendering... 0,784 seconds
Score: 5357

Running 'Database' benchmark...
    Create database... 0,175 seconds
    Populate database... 0,847 seconds
    Save database... 0,117 seconds
    Reload database... 0,045 seconds
    Search database... 0,852 seconds
    Sort database... 0,826 seconds
    Group database... 0,612 seconds
Score: 6187

JRMark (version 22.0.41): 4435

i7-4770T; MSI Z87M Gaming; 8 GB CL9 RAM

=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 4,199 seconds
    Single-threaded floating point math... 2,567 seconds
    Multi-threaded integer math... 1,580 seconds
    Multi-threaded mixed math... 0,804 seconds
Score: 2076

Running 'Image' benchmark...
    Image creation / destruction... 0,355 seconds
    Flood filling... 0,348 seconds
    Direct copying... 0,496 seconds
    Small renders... 1,174 seconds
    Bilinear rendering... 1,056 seconds
    Bicubic rendering... 0,560 seconds
Score: 5516

Running 'Database' benchmark...
    Create database... 0,196 seconds
    Populate database... 1,109 seconds
    Save database... 0,343 seconds
    Reload database... 0,133 seconds
    Search database... 0,916 seconds
    Sort database... 0,928 seconds
    Group database... 0,650 seconds
Score: 5030

JRMark (version 22.0.41): 4207
Logged
WS (AMD Ryzen 7 5700G, 32 GB DDR4-3200, 2x2 TB SDD, LG 34UC98-W)-USB|ADI-2 DAC FS|Canton AM5 - File Server (i3-3.9 GHz, 16GB ECC DDR4-2400, 46 TB disk space) - Media Renderer (i3-3.8 GHz, 8GB DDR4-2133, GTX 960)-USB|Devialet D220 Pro|Audeze LCD 2|B&W 804S|LG 4K OLED )

defenz0r

  • Guest
Re: JRMark -- JRiver's Benchmark Tool
« Reply #346 on: February 11, 2017, 12:36:55 pm »

i7 7700k @4.7 GHz
DDR4 3000 CL 5 RAM
840 EVO 128GB SSD

Can you please add me in the best list? :D

Edit got a better 24/7 OC at 1.20v xD

Probably could score over 7000 if i would oc more but i prefer good voltages and temps.


Faster disks / RAM disk doesn't seem to improve anything.
=== Benchmark läuft (bitte nicht unterbrechen) ===

Benchmark 'Mathematik' wird ausgeführt...
    Single-Thread Ganzzahlmathematik... 3,242 Sekunden
    Single-Thread Gleitkommamathematik... 2,238 Sekunden
    Multi-Thread Ganzzahlmathematik... 1,143 Sekunden
    Multi-Thread gemischte Mathematik... 0,746 Sekunden
Score: 2579

Benchmark 'Bild' wird ausgeführt...
    Bild-Erstellung / -abbau... 0,349 Sekunden
    Flood filling... 0,164 Sekunden
    Direktes Kopieren... 0,314 Sekunden
    Small rendering... 0,813 Sekunden
    Bilinear rendering... 0,664 Sekunden
    Bicubic rendering... 0,359 Sekunden
Score: 8260

Benchmark 'Datenbank' wird ausgeführt...
    Datenbank erstellen... 0,182 Sekunden
    Datenbank befüllen... 0,721 Sekunden
    Datenbank speichern... 0,152 Sekunden
    Datenbank neu laden... 0,061 Sekunden
    Datenbank durchsuchen... 0,679 Sekunden
    Datenbank sortieren... 0,619 Sekunden
    Datenbank gruppieren... 0,490 Sekunden
Score: 7405

JRMark (version 22.0.71): 6081
Logged

Robert van Diggele

  • Recent member
  • *
  • Posts: 10
Re: JRMark -- JRiver's Benchmark Tool
« Reply #347 on: April 09, 2017, 03:54:41 am »

Current score with an i3-6100T

=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 4,808 seconds
    Single-threaded floating point math... 3,316 seconds
    Multi-threaded integer math... 2,450 seconds
    Multi-threaded mixed math... 1,695 seconds
Score: 1549

Running 'Image' benchmark...
    Image creation / destruction... 0,375 seconds
    Flood filling... 0,234 seconds
    Direct copying... 0,418 seconds
    Small renders... 1,236 seconds
    Bilinear rendering... 1,491 seconds
    Bicubic rendering... 0,906 seconds
Score: 4721

Running 'Database' benchmark...
    Create database... 0,206 seconds
    Populate database... 1,163 seconds
    Save database... 0,316 seconds
    Reload database... 0,054 seconds
    Search database... 1,015 seconds
    Sort database... 0,946 seconds
    Group database... 0,744 seconds
Score: 4839

JRMark (version 22.0.88): 3703

The idea is to use this pc for stereo and multichannel flac (mostly 16/44.1 and 24/96) and dsd (highest 128Fs ) files into a USB mch dac.

Plus playback of movies as .mkv with video out on HDMI and audio out on said USB. Audio in those mkv is mostly DTS-MA, so some decoding has to be performed.

Will this do, or do I need a higher CPU?
Thanks.
Logged

JimH

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 71619
  • Where did I put my teeth?
Re: JRMark -- JRiver's Benchmark Tool
« Reply #348 on: April 09, 2017, 06:21:35 am »

Probably.  Please use another thread if you have more questions.
Logged

Robert van Diggele

  • Recent member
  • *
  • Posts: 10
Re: JRMark -- JRiver's Benchmark Tool
« Reply #349 on: April 09, 2017, 09:21:32 am »

Thanks, will do.
Logged
Pages: 1 ... 3 4 5 6 [7] 8 9   Go Up