INTERACT FORUM
Devices => PC's and Other Hardware => Topic started by: robertisonline on December 20, 2015, 09:07:34 pm
-
My mid-2013 MacBook Air is maxed out with a Core i7, 8GB RAM, and PCIe-based flash memory (read/write ~ 720/650, respectively).
=== Running Benchmarks (please do not interrupt) ===
Running 'Math' benchmark...
Single-threaded integer math... 5.483 seconds
Single-threaded floating point math... 3.506 seconds
Multi-threaded integer math... 2.701 seconds
Multi-threaded mixed math... 1.746 seconds
Score: 1414
Running 'Image' benchmark...
Image creation / destruction... 0.625 seconds
Flood filling... 0.541 seconds
Direct copying... 1.457 seconds
Small renders... 1.507 seconds
Bilinear rendering... 1.676 seconds
Bicubic rendering... 0.992 seconds
Score: 3236
Running 'Database' benchmark...
Create database... 0.339 seconds
Populate database... 1.494 seconds
Save database... 0.184 seconds
Reload database... 0.075 seconds
Search database... 1.502 seconds
Sort database... 1.081 seconds
Group database... 0.833 seconds
Score: 3904
JRMark (version 20.0.94): 2851
My Mac Pro is running a quad-core Xeon clocked much higher at 3.7, 10MB L3 Cache, and hyper-threading (I know, I know), 16GB 1866MHz ECC RAM, PCI-e flash with read/write ~ 1400/1400 and of course Dual AMD FirePro's.
JRMark (version 20.0.94): ~3600
So what's exactly the point of JRMark?
-
JMark is a benchmark like any other: It rates how fast a machine can do MC type tasks overall.
Are you confused because your Mac Pro is older, yet it has a higher score? You could compare individual section scores and see which one or ones are better for the Mac Pro. I'm guessing the floating point performance is the big difference. But that's just a guess.
Brian.
-
My MacBook Air is the late-2013 model. I just bought the Mac Pro yesterday. Apple hasn't updated the Mac Pro since late 2013, but I am still surprised that my Mac Pro didn't score much higher than my MacBook Air in JRMark. I am going to return the Mac Pro and get Apple's 27-inch 5K Retina model (late 2015). It is the only computer Apple is currently selling that is using Intel's 6th Gen. Skylake processors. The Mac Pro's are using 3rd gen. Ivy Bridge Xeon processors, which is ridiculous. Spending $3600 on a computer that doesn't' come with a keyboard, mouse, or monitor should be using at least 5th gen Xeon chips and DDR4 RAM. Besides the Mac Pro's dual AMD video cards, Apple's 5K iMac is outperforming the Mac Pro on both single-threaded and multi-threaded benchmarks, even the 6-core Xeon chips. The read/write on the Mac Pro PCIe flash storage is around 1300MBps. Apple's MacBook Pro laptops are approaching 2000 and I'm sure the latest iMac's are doing about the same. The CPU, RAM, and flash storage are all user configurable, but the logic board on the Mac Pro is too old, still using old LGA Intel sockets. None of the JMark scores were above 4300 for the Mac Pro. I'm interested to see how the iMac's i7-6700 with a higher clock speed at 4.2 and faster RAM do on JRMark!
-
My MacBook Air is the late-2013 model. I just bought the Mac Pro yesterday. Apple hasn't updated the Mac Pro since late 2013, but I am still surprised that my Mac Pro didn't score much higher than my MacBook Air in JRMark. I am going to return the Mac Pro and get Apple's 27-inch 5K Retina model (late 2015). It is the only computer Apple is currently selling that is using Intel's 6th Gen. Skylake processors. The Mac Pro's are using 3rd gen. Ivy Bridge Xeon processors, which is ridiculous. Spending $3600 on a computer that doesn't' come with a keyboard, mouse, or monitor should be using at least 5th gen Xeon chips and DDR4 RAM. Besides the Mac Pro's dual AMD video cards, Apple's 5K iMac is outperforming the Mac Pro on both single-threaded and multi-threaded benchmarks, even the 6-core Xeon chips. The read/write on the Mac Pro PCIe flash storage is around 1300MBps. Apple's MacBook Pro laptops are approaching 2000 and I'm sure the latest iMac's are doing about the same. The CPU, RAM, and flash storage are all user configurable, but the logic board on the Mac Pro is too old, still using old LGA Intel sockets. None of the JMark scores were above 4300 for the Mac Pro. I'm interested to see how the iMac's i7-6700 with a higher clock speed at 4.2 and faster RAM do on JRMark!
Four important things to be aware of:
1) Crossplatform JRMarks aren't comparable: running JRiver for Windows and JRiver for Linux on the exact same machine, I see a JRMark in the 5.5K range on windows and a score of about 3.8K on Linux. To provide an even more extreme example, I have JRiver running in a windows virtual machine on a linux server in my attic: the VM Windows JRiver instance actually has a higher JRMarks than the Linux JRiver instance running baremetal on Linux! In fact my windows VM has a higher JRMark than your Mac Pro and I can promise you it's not actually more performant in any ordinary sense of the term. So if you're unhappy with your numbers be sure you're not comparing apples to oranges (or apples to windows as the case may be).
2) JRMarks aren't encessarily comparable over time. As the devs optimize the benchmark, the same PC's score will gradually increase or (in some cases) decrease. It's not designed to be a reproducible benchmark, just a rough order of magnitude of how well a system can perform JRiver tasks.
3) Graphics cards don't really affect the benchmark in any meaningful way. My JRMark didn't change meaningfully when switching over from intel integrated graphics to an NVidia 970 GTX. To the extent GPU's are implicated in the benchmark, it's not a major bottleneck.
4) The single biggest contributor to total JRMark is disk access speed because it will dramatically speed up both the second and third JRMark components. The Processor is not as strong of a contributor. For example, I managed to pick up close to a thousand points by moving an installation from an older sata SSD to a modern pci-e ssd. Processor speed is a secondary driver.
-
=== Running Benchmarks (please do not interrupt) ===
Running 'Math' benchmark...
Single-threaded integer math... 4.010 seconds
Single-threaded floating point math... 2.445 seconds
Multi-threaded integer math... 1.497 seconds
Multi-threaded mixed math... 0.907 seconds
Score: 2145
Running 'Image' benchmark...
Image creation / destruction... 0.377 seconds
Flood filling... 0.330 seconds
Direct copying... 0.266 seconds
Small renders... 1.101 seconds
Bilinear rendering... 0.799 seconds
Bicubic rendering... 0.447 seconds
Score: 6625
Running 'Database' benchmark...
Create database... 0.191 seconds
Populate database... 1.015 seconds
Save database... 0.248 seconds
Reload database... 0.045 seconds
Search database... 0.813 seconds
Sort database... 0.979 seconds
Group database... 0.674 seconds
Score: 5425
JRMark (version 21.0.23): 4732
Media Center
Version: 21.0.23 Registered
JRMark: 4732 (version 21.0.23)
Advanced System Info
Memory & CPU
CPU features: MMX, SSE2, SSE3, SSSE3, SSE4.1, AVX
Free Physical Memory: 12.8 GB
Total Memory: 15.9 GB
Operating System
Microsoft Windows 10 64-bit
Audio inputs and outputs
AMD HDMI Output (AMD High Definition Audio Device) (driver 10.0.10586.0)
Realtek Digital Output (Realtek High Definition Audio) (driver 10.0.10586.0)
Realtek Digital Output(Optical) (Realtek High Definition Audio) (driver 10.0.10586.0)
Speakers (Realtek High Definition Audio) (driver 10.0.10586.0)
DiskDrive
Samsung SSD 850 EVO 250GB (driver 10.0.10586.0)
Display
AMD Radeon HD 5800 Series (driver 15.201.1151.1008)
MEDIA
AMD High Definition Audio Device (driver 10.0.0.1)
Realtek High Definition Audio (driver 6.0.1.7709)
Processors
Intel(R) Core(TM) i7-5930K CPU @ 3.50GHz (driver 10.0.10586.0)
Intel(R) Core(TM) i7-5930K CPU @ 3.50GHz (driver 10.0.10586.0)
Intel(R) Core(TM) i7-5930K CPU @ 3.50GHz (driver 10.0.10586.0)
Intel(R) Core(TM) i7-5930K CPU @ 3.50GHz (driver 10.0.10586.0)
Intel(R) Core(TM) i7-5930K CPU @ 3.50GHz (driver 10.0.10586.0)
Intel(R) Core(TM) i7-5930K CPU @ 3.50GHz (driver 10.0.10586.0)
System
Intel(R) C610 series/X99 chipset SMBus Controller - 8D22 (driver 10.1.2.10)
Intel(R) C610 series/X99 chipset SPSR - 8D7C (driver 10.1.2.10)
Thought I would post as a reference design for other looking to build a multi-purpose htpc. Not mentioned/obvious in sysinfo: ASUS X99-M WS motherboard, Corsair CMK16GX4M4A2133C13 DDR4
-
=== Running Benchmarks (please do not interrupt) ===
Running 'Math' benchmark...
Single-threaded integer math... 3,530 seconds
Single-threaded floating point math... 2,154 seconds
Multi-threaded integer math... 1,186 seconds
Multi-threaded mixed math... 0,695 seconds
Score: 2512
Running 'Image' benchmark...
Image creation / destruction... 0,316 seconds
Flood filling... 0,360 seconds
Direct copying... 0,586 seconds
Small renders... 0,912 seconds
Bilinear rendering... 0,651 seconds
Bicubic rendering... 0,395 seconds
Score: 6833
Running 'Database' benchmark...
Create database... 0,089 seconds
Populate database... 0,890 seconds
Save database... 0,246 seconds
Reload database... 0,162 seconds
Search database... 1,223 seconds
Sort database... 0,689 seconds
Group database... 0,677 seconds
Score: 5408
JRMark (version 26.0.56 64 bit): 4918
-----
Media Center
Version: 26.0.56 registered
Install path: C:\Program Files\J River\Media Center 26\
Interface plug-ins: Interface Plugins: TiVo Server (not running)
JRMark: 4918 (version 26.0.56)
Program run time: 23,2 hours
Memory used: 73,1 MB memory
Handles used: 802 handles
Advanced System Info
Memory & CPU
CPU features: MMX, SSE2, SSE3, SSSE3, SSE4.1, AVX
Free Physical Memory: 16,7 GB
Total Logical Cores: 8
Total Memory: 23,9 GB
Operating System
Microsoft Windows 10 64-bit
Disk drives
KINGSTON SA400S37120G (driver 10.0.18362.1)
Samsung SSD 850 EVO 250GB (driver 10.0.18362.1)
Samsung SSD 860 EVO 2TB (driver 10.0.18362.1)
ST4000DM000-1F2168 (driver 10.0.18362.1)
Display adapters
Intel(R) HD Graphics 4600 (driver 20.19.15.5107)
Processors
Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz (driver 10.0.18362.693)