I have a few Nvidia-based graphics boards that aren't on your list yet.

This thread is rather long to search through for restrictions on what else can be running on the computer during these tests, so could one of you summarize them?

The latest requirements are in my last post, 3 posts up! ;)
As I said in my 1st post I would add updated requirements & a benchmark table towards the end of the thread.:)
.... wait, I seen I've already answered you in my last post! Are you actually reading anything?? :p
It would be good to have more Nvidia cards on the list btw.
.
Todderbert
Thanks for your time :), I will add it to the AnandTech table today, & add an updated table here when I have a few more times from others.

---

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

Happy to see a new benchmark started, since I had to replace the card used in the previous test because one of the two fans seized. So I replaced the 7950 TOP with an R9 280X TOP (ASUS). It has a factory OC of 1070Mhz and 1600Mhz on the memory. I use the latest Boinc version and AMD graphics driver (not the beta) on Win7 64bit Pro. Run times to complete "106.88 credit WU's" average in 25sec range and I provided a link for reference. Regards, Wes.

http://milkyway.cs.rpi.edu/milkyway/results.php?userid=1017038&offset=0&show_names=0&state=4&appid=

You have mail!!

My GTX Titan X stock: 101.1s
My GTX 960 SC (GPU 1366.9mhz): 253.7s

Just an FYI that I don't notice any single task speed difference between running just 1 or 5 tasks simultaneously per GPU on my Titan Black cards, nor do I notice any difference between using one CPU thread per GPU task or 0.09 CPU threads per GPU task. With Double Precision optimization enabled in the Nvidia Control Panel, 5 tasks running simultaneously per GPU only loads them about 79% and only loads the VRAM about 36% above background usage (43% total) while using SLI – I also game on this computer, host game servers on a GNU/Linux (Debian "Jessie" 64bit/KDE) VM, et cetera and typically run 4 tasks simultaneously per GPU (8 total) and 8 tasks simultaneously on the CPU. Using SLI essentially doubles VRAM usage as it mirrors memory on both GPUs, even though BOINC doesn't take advantage of SLI (nor would it really make any sense for it to), so it is not recommended to use SLI unless you're using it for something else and have more than enough VRAM to play around with as I do.

I'm using an i7-3930K with all cores clocked to 4.2GHz and the OS I'm using is Windows 7 Pro 64bit SP1.

Titan Black GPUs are clocked to 966MHz, and as mentioned, Double Precision optimization is enabled in the Nvidia Control Panel.

The current GPU MilkyWay@Home 1.02 (opencl_nvidia) application tasks take about 00:02:25 to complete...

...and the current GPU MilkyWay@Home Separation (Modified Fit) 1.36 (opencl_nvidia_101) application tasks take about 00:00:35 to complete.

---

...

Update:

Completion times are identical with Double Precision optimization disabled in the Nvidia Control Panel while running 1 GPU tasks per GPU and CPU thread – 00:02:25 (145s) and 00:00:35 (35s) respectively.

However... these same kinds of tasks load the GPU to about 3 to 4 times more when Double Precision optimization is disabled! This means I can only effectively run 1 or maybe 2 tasks simultaneously per GPU before the GPU load capacity is full and significantly bottlenecks performance. Coincidentally, maxing out the load on the GPUs this way somewhat strangely also seems to spill the workload over onto the CPU.

And again... these are all on my Titan Black cards (EVGA GeForce GTX Titan Black Superclocked). The "superclocked" Titan Blacks basically just have a very slightly higher base clock and boost clock rate, which is effectively irrelevant when using Double Precision optimization as it locks in the clock speed to 966MHz while raising GPU voltage a little, though as mentioned I'm getting identical completion times with Double Precision optimization disabled and the GPUs boosting up a little anyway.

---

Hi, I just started crunching for MW@Home -

Computer 690635:

CPU: Intel Pentium D CPU 2.80GHz (stock)
OS: Microsoft Windows XP Media Center x86 Edition, Service Pack 3
BOINC version: 7.6.22
Memory: 3838.09 MB
Cache: 1024 KB
Application: MilkyWay@Home v1.00

22 valids:

Average Run Time: 9946.205 sec.
Average CPU Time: 9937.431 sec.

Looks like I am slow man on the totem pole!

Andrew :D

---

...

Update 2:

It had been a little while since I updated Windows and my Nvidia drives and since I had restarted this computer – about a month. So, I did those and performed some more tests to help eliminate any potential performance influencing factors. However, nothing really changed.

Even with the updated Nvidia drivers, SLI disable, and not running a VM for my game servers nor running any additional separate BOINC CPU tasks, the results remain identical to my former results with Double Precision optimization disable and enabled.

The GPU tasks I'm running that I mentioned are the 160.88 and 20 credit tasks. I believe anyone is able to check this host's listed results to confirm this as they please.

I can only assume that there is some flaw in the methodology I or others are using who have tested and submitted results for Titan Black cards.

Cheers.

---

...

Welcome, and thanks for your efforts!

With distributed computing we can all do our parts and every little bit helps.

Some of us have pretty ridiculous computer systems that we've spent many thousands of dollars on. These computers are an exception though, and I would presume that the majority of work is actually done on more modest systems.

---

...

The GPU tasks I'm running that I mentioned are the 160.88* and 20 credit tasks. I believe anyone is able to check this host's listed results to confirm this as they please.

...

Correction: *106.88

Also, here are the validated host results for this computer, if anyone want's to check... http://milkyway.cs.rpi.edu/milkyway/results.php?hostid=478184&offset=0&show_names=0&state=4&appid=

Don't get me wrong, I wouldn't mind if it only took 24 seconds to crunch these work units on my Titan Black cards. ;)

---

Numbers are very slightly slower (about an additional second) with the additional CPU tasks I'm currently running, it seems, and in case anything else changes in the meantime that would throw off the results in the link I posted, here's a picture instead that's more relevant to this specific test...

https://i.imgur.com/HgG7EJu.png

I now return you to your regularly scheduled thread... (and sorry for the post bombing – I wish I could edit posts over an hour old instead).

---

After a hiatus from MilkyWay I am back, and I have dedicated two 7970s to crunch on MilkyWay until they die. ;)

OS: Windows 7 x64
BOINC Version: 7.6.22
GPU Driver: 14.9 WHQL
CPU: AMD FX8320e @ 4.8 GHz
GPU: AMD HD7970 @ 1250/1550

20 WU avg run-time - 21.02s

Running two tasks concurrently has slightly better throughput, with each task taking an average of 35.5 seconds to complete.

I tested the following on a host with a G1610 @ 2.60GHz running Linux with driver 355.11

GTX 650 Ti (stock) - Average Run: 519.6 sec, Average CPU: 18.4 sec.

Tesla M2090 (stock) - Average Run: 84.2 sec, Average CPU: 77.5 sec.

I will retry the Tesla with a faster CPU to see if the processor was impacting performance. I would have expected it to be somewhere between a 5850 and a 7950 based on its DP rating.

Sapphire R9 380 here with stock clocks [1010/1450], 96.72s average runtime on Win7 x64 (and a C2D E6750^^)

Hi folks, sorry I've not been around, time for an update!
Thanks for your input & times :).
.
Wrend
No flaw in the methodology, I see you've mis-understood how the Titan's times are derived, which on re-reading it I can see why, it isn't clear so I will edit the explanation.
Anyway the 23 & 25s times are derived from the total run time of running 'x' WUs at once divided by the number of WUs. So in these 2 cases they ran 8 WUs at once (seems to be a favourite choice), so their total times were divided by 8.
It's not as accurate as running 1 WU at a time as per the AMD cards, but it's the only way to properly load them up.
Leave DP optimisation on btw.

Regarding leaving a free CPU core for each GPU, that does work when the GPU can be fully loaded in the 1st place, so for AMD cards leaving a free CPU for each GPU does improve GPU WU times.

And yea I hear ya on the pointless 1hr edit limit!

Todderbert
How many WUs were you running on your Titan X at the time? 8? (& the 960 for that matter).

SharkNose
Old skool a? ;), I always like to see how old gear runs modern DC projects, thanks for the time :), will add it shortly

tictoc
Thanks for your time, a new record! :)

fractal
Your tesla's time is probably suffering due to the same problem as other Nvidia cards, of MW with Nvidia cards not fully loading the GPU. Check your GPUs load.
Try running multiple WUs at once.

(I'm wondering whether to allow all Nvidia cards to run multiple WUs for the benchmark table..........)

andre
Thanks for your time :), will add it shortly.
How many WUs was that time derived from btw?

**************************************************

Will post a new table here when the above questions have been answered.
I have updated the AnandTech thread http://forums.anandtech.com/showthread.php?t=2384985

---

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

Your tesla's time is probably suffering due to the same problem as other Nvidia cards, of MW with Nvidia cards not fully loading the GPU. Check your GPUs load.
Try running multiple WUs at once.

(I'm wondering whether to allow all Nvidia cards to run multiple WUs for the benchmark table..........)

I can try running multiple units but it looks like the card is fully loaded as it is

Thu Jun  2 18:02:52 2016
+------------------------------------------------------+
| NVIDIA-SMI 355.11     Driver Version: 355.11         |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  Tesla M2090         Off  | 0000:01:00.0     Off |                    0 |
| N/A   N/A    P0   167W / 225W |    220MiB /  5375MiB |     99%      Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID  Type  Process name                               Usage      |
|=============================================================================|
|    0      2008    C   ...n_1.02_x86_64-pc-linux-gnu__opencl_nvidia   209MiB |
+-----------------------------------------------------------------------------+

edit: Running two work units at a time shaves 9 seconds off the run time and keeps both CPU cores fully utilized but does not have any appreciable affect on the Tesla.

Thu Jun  2 18:21:18 2016
+------------------------------------------------------+
| NVIDIA-SMI 355.11     Driver Version: 355.11         |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  Tesla M2090         Off  | 0000:01:00.0     Off |                    0 |
| N/A   N/A    P0   168W / 225W |    431MiB /  5375MiB |     99%      Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID  Type  Process name                               Usage      |
|=============================================================================|
|    0      2356    C   ...n_1.02_x86_64-pc-linux-gnu__opencl_nvidia   209MiB |
|    0      2364    C   ...n_1.02_x86_64-pc-linux-gnu__opencl_nvidia   209MiB |
+-----------------------------------------------------------------------------+

OS: Windows 10 64-bit
BOINC Version: 7.6.33
GPU Driver: 16.12.2
CPU: Intel i5-6600K 4.6 GHz
GPU: AMD RX 480 1415 MHz Core/2025 MHz Memory

5 WU Average Run Time: 72.128 sec

Major change, at some point last autumn(ish) the old MW v1.0x app & it's WU were replaced with a new app & new WUs.
So currently there is only 1 WU credit type, of 133.66 credits.

This of course means an entire new benchmark table!

*******************************************************************************

Thanks for your time Todd :)
Just to confirm, that's with 133.66 credit WUs right?
Will add your time to the table & post it here.

Fractal
Hi, sorry I didn't reply to your post earlier, curious to see 99% util running 1 WU, but still a performance improvement running 2!

---

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

New benchmark table & requirements for the 133.66 credit WU

So please share your new scores for old & new GPUs & CPUs alike!




Note new requirements for the benchmark :-




Please use validated 133.66 credit WU results only, they must be from the MilkyWay@Home v1.4x app

(Currently for the GPUs, there is only 1 app)




Average of at least 5 WU times (not cherry picked please! ;)).




A dedicated physical CPU core for each GPU (for optimal MW WU times). If only using BOINC for CPU tasks, & you have an HT capable CPU, then the only way to be certain of this (bar disabling HT) is to set the BOINC computing preferences (in advanced mode>options) so that you have 1 less CPU thread running then you do physical cores. Don't panic too much about lost CPU ppd, it doesn't take long to run MW GPU WUs ;) (see table).

Please state what speed & type CPU you have, as it now seems to have a significant affect on GPU WU times!




Please state GPU & RAM clock speeds if overclocked (including factory overclocks) or state 'stock'.




Please only crunch 1 WU at a time per GPU, otherwise it will massively increase WU time! (even if it does increase output, the WU times seem to fluctuate much more than singly crunched WUs, so you can't just 1/2 the times either).

[note, the following paragraph may no longer be relevant for v1.4x, time will tell] I've decided to relent a bit on this, but only for the GTX Titan as it can't achieve anywhere near full load with just 1 WU, I will add a proviso stating this by each Titan's score (which will be derived from the total time crunched, divided by the number of WUs being crunched simultaneously. 8 WUs at once seems to be the choice so far).




For CPU times please state whether Hyper Threading (or equivalent) is enabled or not, times for both states welcomed :).




It would also be useful if you could state your BOINC & driver version, & OS, incase it does make any difference.




[This paragraph under review]

If you find your WU times are fluctuating more than a couple of % then use GPU-Z or your grx card driver tools to check that you GPU is able to hit near 100% load (although I'm not sure that NVidia cards can hit that for MW), note that even when crunching normally, the GPU load will be on/off on this current MW app, so the GPU load graph should look like a series of blocks.

Also check using task manager that your CPU does actually have the spare load to give to MW (& btw, GPU crunching won't show up in the TM).




***************************************************************************************************************




Current GPU statistics ~ Average Run Time to Complete 1 MW v1.4x 133.66 credit WU :-




R9 280X, GPU 1080 MHz (CPU, Pentium G3220 @3 GHz) ................................. 40.1s ... Tennessee Tony

HD 7970, GPU 1000 MHz (CPU, i7 4930k @4.1 GHz) ......................................... 42s ...... Assimilator1

R9 280X, Stock? (CPU, C2D E6550, stock?) .......................................................... 54.3s ... iwajabitw

R9 280X, GPU 1020 MHz (CPU, AMD FX8320E @3.47 GHz) .............................. 54.8s ... Tennessee Tony

HD 7950, GPU 860 MHz (CPU, i7 3770k, stock) .................................................. 56.5s ... salvorhardin

HD 7870 XT 3GB(DS), GPU 925 MHz (CPU, C2 Q9550 @3.58 GHz) .................. 56.8s ... Assimilator1

R9 390, GPU 1015 MHz (CPU, i7 3770k, stock) ................................................... 60.7s ... salvorhardin

RX 480, GPU 1415 MHz, RAM 2025 MHz (CPU, i5 6600k, 4.6 GHz) .................. 72.1s ... TomTheMetalGod

HD 6950, stock (CPU Athlon2 X4 620 @2.6 GHz) ............................................. 101.2s ... waffleironhead

GTX 980, GPU 1303 MHz (CPU, i7 5820k @3.3 GHz) ........................................ 184s ...... iwajabitw

Quadro K2100M, stock (CPU, i7 4900 MQ turbo @3.8 GHz) ......................... 1784s ...... StefanR5R




Current CPU statistics ~ Average Run Time to Complete 1 MW v1.4x 133.66 credit WU :-




i7 5820k @3.3 GHz ......................................................................... 2723s no 'HT load' .... iwajabitw

i7 4930k @4.1 GHz (6 threads for CPU) ....................................... 2825s no 'HT load' ... Assimilator1

i7 4930k @4.1 GHz (10 threads for CPU, 2 for GPU)................... 4171s HT on .............. Assimilator1

I7 4930k @4.1 GHz (12 threads for CPU) ..................................... 4557s HT on .............. Assimilator1




****************************************************************************************************************

Info:-




My previous MW benchmark thread spring 2014 - summer 2016




Stock clocks for some of the commonly used graphics cards for MW (& cards with good double precision power), source Wiki (GPU/RAM MHz or MT/s if stated) :-

AMD .............................GPU/RAM ................................... DP GFLOPS

HD 4890 ...................... 850/975 ....................................... 272*

HD 5830 ...................... 800/1000 ..................................... 358

HD 5850 ...................... 725/1000 ..................................... 418

HD 5870 ...................... 850/1200 ..................................... 544

HD 5970 ...................... 725/1000 (dual GPU) .................. 928

HD 6930 ...................... 750/1200 ..................................... 480

HD 6950 ...................... 800/1250 ..................................... 563

HD 6970 ...................... 880/1375 ..................................... 675

HD 6990 ...................... 830/1250 (dual GPU) ................ 1277

HD 7870 XT ................. 925-975/1500 ............................. 710-749

HD 7950 ...................... 800/1250 ..................................... 717

HD 7950 Boost ........... 850-925/1250 ............................. 762-829

HD 7970 ...................... 925/1375 ..................................... 947

HD 7970 GE ............... 1000-1050/1500 ......................... 1024-1075

HD 7990 ..................... 950-1000/1500 (dual GPU) ....... 1894-2048

R9 280 ........................ 827-933/1250 .............................. 741-836

R9 280X ...................... 850-1000/1500 ............................ 870-1024

R9 290 ........................ >947/5000 MT/s .......................... 606

R9 290X ...................... >1000/5000 MT/s ....................... 704

R9 295X2 .................... 1018/5000 MT/s (dual GPU) .... 1433

R9 390 ........................ >1000/6000 MT/s ....................... 640

R9 390X ...................... >1050/6000 MT/s ....................... 739

R9 Fury ....................... 1000/1000 MT/s ......................... 448

R9 Nano ..................... 1000/1000 MT/s .......................... 512

R9 FuryX .................... 1050/1000 MT/s .......................... 538

R9 Pro Duo ................ 1000/1000 MT/s (dual GPU) ...... 900

RX 470 ........................ 926-1206/6600 MT/s .................. 237

RX 480 ...................... 1120-1266/7000-8000 MT/s ........ 323




Wow, just noticed how feeble the entire R 400s line is at Double Precision!, even the top of the line (as of 12/16) RX 480 only manages 323 GFLOPs, which is a little less than the HD 5830s 358 from 2/10 & only a bit more than the HD 4890 from 4/09! Although it is more than the R9 380X's 248 GFLOPs :p.




I can see it won't be long before we have ancient 5800s, 6900s & 7900s (& 7870 XTs) as a secondary card in our rigs solely for crunching MW & Einstein, & modern cards for gaming & SP DC! ..........maybe I'm behind the times & some of you guys are already doing that!? ;)




* The 4800s can't run MW atm, see here




NVidia ...............................GPU/RAM ....................... DP GFLOPS

GTX 980 ................ 1126-1216 MHz/7010 MT/s .............. 144

GTX 1070 .............. 1506-1683 MHz/8000 MT/s .............. 181-202

GTX 1080 .............. 1607-1733 MHz/10,000 MT/s ........... 257-277

New benchmark thread at the AnandTech forums (is updateable!) https://forums.anandtech.com/threads/milkyway-h-benchmark-thread-winter-2016-on-different-wu-gpu-cpu-times-wanted.2495905/#post-38654842 .

And of course the rubbish forum software here completely balls up the formatting & links!
Checkout the AnandTech thread for working links.

---

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

OS: Windows 10 pro 64bit
BOINC Version: 7.6.33 64
GPU Driver: 17.1.1 no whql
CPU: intel I7 5960X @ 4.0 mhz HT off (pga primegrid LLR) (2133ddr4mhz 64gb)
GPU 1: AMD HD7970 MSI R7970 LIGHTNING @ 925/1375 (default)
GPU 2: AMD HD7970 Vapor-X HD 7970 GHz Edition 3G D5 @ 1050/1500 (default)
Host ID:718384
https://milkyway.cs.rpi.edu/milkyway/results.php?hostid=718384&offset=0&show_names=0&state=4&appid=
Time : 39.xx and 41.xx sec.
Stable and not cherry picked results. On default stock gpu mhz s..
39 second :
https://milkyway.cs.rpi.edu/milkyway/result.php?resultid=1949021186

I will later add overclocking but stable results. Becouse i have there "big" overclock options on cpu and gpu ..))