I latter realised for the CPU tests only that I should of used & asked for the 'CPU time' rather than 'run time' which would include any slow downs due to running other processes. I think that largely due to the use of averaging at least 5 WU times that it will keep the inaccuracy low & that the times are still useful, just bear in mind their will be a little inaccuracy there. Just re-iterate that this refers to the CPU benchmarks only.

I've just checked through 1100 of my valid results & their were only 3 results with long times but I think those were where I down clocked the GPU (for YouTube!) or where I was watching a video.

So I think it's safe to say now the 'long' 213.76 credit WUs are extinct, this benchmarking run is now over.
Oh & my RAC has now gone upto 123k thanks to the short 213s! :)

At some point I'll do another benchmarking thread, this time using the MW v1.02 WUs. No plans as to when yet.

---

Team AnandTech - SETI@H, DPAD, F@H, MW@H, A@H, LHC, POGS, R@H, Einstein@H, DHEP, WCG

Main rig - Ryzen 5 3600, MSI B450 G.Pro C. AC, RTX 3060Ti 8GB, 32GB DDR4 3200, Win 10 64bit
2nd rig - i7 4930k @4.1 GHz, HD 7870 XT 3GB(DS), 16GB DDR3 1866, Win7

Well it looks like the long 213.76 credit WUs are back!

So if you have the results, post the times :) (as per the req's in my post 61587).

---

Team AnandTech - SETI@H, DPAD, F@H, MW@H, A@H, LHC, POGS, R@H, Einstein@H, DHEP, WCG

Main rig - Ryzen 5 3600, MSI B450 G.Pro C. AC, RTX 3060Ti 8GB, 32GB DDR4 3200, Win 10 64bit
2nd rig - i7 4930k @4.1 GHz, HD 7870 XT 3GB(DS), 16GB DDR3 1866, Win7

All results under linux with NVidia 331.67 and AMD 14.4-2:

Run time (sec) | CPU time (sec) | Credit
AMD HD 7870 boost (tahiti Chip) | 152.99 | 8.21 | 213.76 (Two WU's simultaneously; so divide Run time by 2.)
83.65 | 40.15 | 106.88 (Two WU's simultaneously; so divide Run time by 2.)

AMD R7 Spectre (A10-7850K) | 1,105.01 | 9.96 | 213.76
553.32 | 5.02 | 106.88

AMD R3 Kalindi (Athlon 5350) | 4,816.01 | 2,056.16 | 213.76
2,411.88 | 1,018.88 | 106.88

NVIDIA GeForce GT 635M | 1,222.49 | 77.63 | 106.88 (Modified Fit don't work, because the libc6 is too old. But you can, as you can see in the other, appreciate the double time: 2,444 s)

Just switched the rig into benching mode (disable HT, remove app_config.xml, reserve 1 core pr gpu in general settings for Boinc).
specs : Windows 7, dual Xeon X5570@2.93 Ghz(turboboost up to 3.33 Ghz/generally running 3.2Ghz turboboost), 48 GB DDR3 ECC ram, dual AMD Radeon R9-280X@1150 mhz core/1600 mhz memory. Catalyst 14.4 drivers

213.76 WU's average time 45.5 seconds
example :
task wu comp date_sent date_submit status run_time cpu_time credit
749533384 560591196 561358 23 May 2014, 15:39:09 UTC 23 May 2014, 16:10:44 UTC Completed and validated 46.09 4.32 213.76

106.88 WU's average time 24.5 seconds.
example :
task wu comp date_sent date_submit status run_time cpu_time credit
749532616 560662815 561358 23 May 2014, 15:38:01 UTC 23 May 2014, 16:10:44 UTC Completed and validated 24.10 2.61 106.88


231.76 WU CPU average time 1 hour, 35 minutes and 4.44 seconds.
example:
749539775 560668133 561358 23 May 2014, 15:51:13 UTC 23 May 2014, 17:29:56 UTC Completed and validated 5,704.44 5,688.75 213.76

My system spec:

CPU Intel Core i7 CPU X 980 @ 3.33GHz
AMD ATI Radeon HD 5970 (gpu clock 735 and memory clock 1010)
Catalyst version 12.6
Windows 7 Ultimate x64 Edition
BOINC version 7.0.64

Only gpu crunching:
time vary from 144 to 150 sec

With full cpu load on other project:
155 to 161 sec.

The HD 5970 is a dual-gpu card and I have a 100% gpu utilization with 1 WU/Core.

Thanks for your scores guys but ........

All results under linux with NVidia 331.67 and AMD 14.4-2:

Run time (sec) | CPU time (sec) | Credit
AMD HD 7870 boost (tahiti Chip) | 152.99 | 8.21 | 213.76 (Two WU's simultaneously; so divide Run time by 2.)
83.65 | 40.15 | 106.88 (Two WU's simultaneously; so divide Run time by 2.)

AMD R7 Spectre (A10-7850K) | 1,105.01 | 9.96 | 213.76
553.32 | 5.02 | 106.88

AMD R3 Kalindi (Athlon 5350) | 4,816.01 | 2,056.16 | 213.76
2,411.88 | 1,018.88 | 106.88

NVIDIA GeForce GT 635M | 1,222.49 | 77.63 | 106.88 (Modified Fit don't work, because the libc6 is too old. But you can, as you can see in the other, appreciate the double time: 2,444 s)

No I can't divide the WU times in 2, because as I mentioned in my benchmark requirements (post 61587) running 2 WUs in parallel on 1 GPU screws up times.
Didn't you read the benchmark requirements?

DutchDK
Looking at your 280X times vs what's already posted you have picked times from the short versions of the 213.76 credit WUs. I did discuss this in the benchmark requirements. Did you not have any long/slow 213.76s?

superpower
The above problem applies to you too, the clock speeds of the 5970 are less than the 5870 so your times should be slower than the 5870 (though 2 WUs at a time of course :) ), you've also used times from the faster version of the 213.76 credit WUs. Did you not have long versions either? (read benchmark requirements in post 61587 for details).

---

Team AnandTech - SETI@H, DPAD, F@H, MW@H, A@H, LHC, POGS, R@H, Einstein@H, DHEP, WCG

Main rig - Ryzen 5 3600, MSI B450 G.Pro C. AC, RTX 3060Ti 8GB, 32GB DDR4 3200, Win 10 64bit
2nd rig - i7 4930k @4.1 GHz, HD 7870 XT 3GB(DS), 16GB DDR3 1866, Win7

Can anyone posting benchmark times please read the benchmark requirements from start to end in message 61587, it's not that long! ;)

Otherwise it's very likely the times aren't relevant & are therefore unusable :(

[update] Well I've just looked through 300 valid results by DutchDK & I didn't see a single long 213.76 WU time, so it looks like they've disappeared again!

So regarding benchmarking, if their are no long 213.76s then the times can't be used for this benchmarking thread, look out for any long ones hidden deep though! ;)

---

Team AnandTech - SETI@H, DPAD, F@H, MW@H, A@H, LHC, POGS, R@H, Einstein@H, DHEP, WCG

Main rig - Ryzen 5 3600, MSI B450 G.Pro C. AC, RTX 3060Ti 8GB, 32GB DDR4 3200, Win 10 64bit
2nd rig - i7 4930k @4.1 GHz, HD 7870 XT 3GB(DS), 16GB DDR3 1866, Win7

Not sure if what i do post is what your looking for...
Atleast my grafic card is new to your list.. :)

PC ID: 547777
BOINC version: 7.2.42
CPU: AMD Phenom(tm) II X4 810 Processor [Family 16 Model 4 Stepping 2] (4 processors)
GPU: [2] NVIDIA GeForce GTX 690 (2048MB) driver: 332.21 OpenCL: 1.01
OS: Microsoft Windows Vista Home Premium x64 Edition, Service Pack 2, (06.00.6002.00)

The app_config.XML file iam using with the normal setup

<app_config>
<app>
<name>milkyway_separation__modified_fit</name>
<gpu_versions>
<gpu_usage>0.25</gpu_usage>
<cpu_usage>0.250</cpu_usage>
</gpu_versions>
</app>
</app_config>

From this app_config file you can see that i do crunch 4x WUs / GPU with 1x CPU / GPU
My grafic card has 2 GPU's
0: GeForce GTX690 @1019,5 Mhz (Stock) / 2Gb Mem @3004 MHz
1: GeForce GTX690 @1019,5 Mhz (Stock) / 2Gb Mem @3004 MHz

So in total iam crunching for MW 8 WUs at once with my grafic card
~Crunch time per 8 WUs = 00:32:00 // 0,25C + 0,25NV (d0) and (d1)
I do use the programm TThrottle with the max temp for the GPUs set @76 Celcius

Task Work unit Computer Run time (sec) CPU time (sec) Credit Application
752798431 563115092 547777 1,896.14 516.27 213.76 Milkyway@Home Separation (Modified Fit) v1.30 (opencl_nvidia)
752785741 563105786 547777 1,906.79 542.84 213.76 Milkyway@Home Separation (Modified Fit) v1.30 (opencl_nvidia)
752785733 563105780 547777 1,891.66 526.69 213.76 Milkyway@Home Separation (Modified Fit) v1.30 (opencl_nvidia)
752784405 563104801 547777 1,886.71 514.65 213.76 Milkyway@Home Separation (Modified Fit) v1.30 (opencl_nvidia)
752781763 563102905 547777 1,871.44 504.79 213.76 Milkyway@Home Separation (Modified Fit) v1.30 (opencl_nvidia)

The average time for crunching 8 WUs at a time is 1890sec

Addon to the above posting:

The results with the request test setup:

<app_config>
<app>
<name>milkyway_separation__modified_fit</name>
<gpu_versions>
<gpu_usage>1</gpu_usage>
<cpu_usage>1</cpu_usage>
</gpu_versions>
</app>
</app_config>

Temperature / GPU = 67 Celcius by 100% usage

Did crunched 1 WU / GPU with 1 CPU / GPU
The GeForce GTX 690 has 2 GPUs on board

Here the last 5 best results:

752863778 563080326 547777 877.27 7.88 213.76
752844114 563147010 547777 824.26 7.94 213.76
752837042 563067613 547777 781.25 7.83 213.76
752834726 563140497 547777 875.29 7.72 213.76
752834725 563140496 547777 839.28 8.05 213.76

The average time for crunching is now ~839sec

Hi JBG
Good to talk to a S@NL member again :) (use to regularly in the SETI classic days).
I would certainly like to add your GPU to the list, unfortunately the times in your 1st post aren't useable (running more than 1 WU/GPU gives widely varying times).

Your method in your 2nd post is correct though :), so I may be able to use those times if they are from the long version of the 213.76 credit WUs. Although your times still seem to be varying a lot.

I will check your WU times to see if they are the short or long 213.76 WUs..........

Hmmm, I can't, I assume you've gone back to crunching 4 WU's/GPU? (213.76 times of ~1600s).

I think using the 213.76 credit WUs is just too difficult now, trying to find the long versions now is at best rare & at worst sometimes non-existant (currently I'm getting a mix of long & short ones).
It's possible to determine on setups always just crunching 1WU/GPU if they are crunching long or short versions but not where more than 1WU/GPU is crunched.

I'm going to start a new benchmark thread using the MW v1.02 WUs (some of the 106.88 credit WUs). I may post it tonight, it would be good to have your times there :), and everyone elses! ;).

---

Team AnandTech - SETI@H, DPAD, F@H, MW@H, A@H, LHC, POGS, R@H, Einstein@H, DHEP, WCG

Main rig - Ryzen 5 3600, MSI B450 G.Pro C. AC, RTX 3060Ti 8GB, 32GB DDR4 3200, Win 10 64bit
2nd rig - i7 4930k @4.1 GHz, HD 7870 XT 3GB(DS), 16GB DDR3 1866, Win7

yep sorry, i just did drop back to 1wu per gpu for the test for a short time..

just 6 days ago i had ~223k credits @MW, and now after the 6 days crunching
with 8 WUs (2x GPU) and 2cpu WUs @24/7 the credit count is raised to ~605k :)
(day 1 to day 6: 43k, 52k, 46k, 51k, 78k, 83k credits)

go do your test next time a bit longer, so that you can do a much better average time calculation.. for my card
will keep an eye on the test topic the next days :D

Edit:
the chrunching time of the MW v1.02 @ 8WUs = approx 10min.. per WU

Please refer to new thread for further discussion.

http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=3551

---