I have not set this up as a 'nice' cron job yet, so this is currently running at a user 'nice' level of 100%. Seems to average around 70% to 95% of CPU use.
Fedora Core 25, stock Dell Precision 390 - no GPUs
Running between 70 to 90% of CPU
Intel(R) Core(TM)2 CPU 6700 @ 2.66GHz [Family 6 Model 15 Stepping 6]
Linux, 4.9.5-200.fc25.i686
boinc: 7.6.22

Numbers below were running two at a time, 1 per CPU core.
-----------

1955052971 1420541770 718326 31 Jan 2017, 1:02:11 UTC 2 Feb 2017, 22:57:51 UTC Completed and validated 6395.83 4762.3 133.66 MilkyWay@Home v1.40
1955052975 1420564855 718326 31 Jan 2017, 1:02:11 UTC 2 Feb 2017, 20:09:28 UTC Completed and validated 7674.7 4758.43 133.66 MilkyWay@Home v1.40
1955052978 1420578011 718326 31 Jan 2017, 1:02:11 UTC 2 Feb 2017, 22:57:51 UTC Completed and validated 6421.73 4768.28 133.66 MilkyWay@Home v1.40
1955052979 1420578070 718326 31 Jan 2017, 1:02:11 UTC 2 Feb 2017, 21:38:36 UTC Completed and validated 6971.88 4879 133.66 MilkyWay@Home v1.40
1955053015 1420599137 718326 31 Jan 2017, 1:02:11 UTC 2 Feb 2017, 21:38:36 UTC Completed, validation inconclusive 6657.22 4745.91 pending MilkyWay@Home v1.40

---

http://www.joescat.com/boinc/

Windows 7 Ultimate 64 bit - BOINC 7.6.33 - GPU Driver 14.1
Xeon E3 1270 V3 @ 3.7 GHz HT on
MW 1 task only - GPU @ 1065MHz - Memory @ 1555MHz 40 seconds - NO CPU tasks !
Same with CPU tasks limited to 88% - 43 seconds

dunx

P.S. Tried two tasks per GPU, and it messed up all four GPUs.... liking the sensible VRM temps though !

Looks like I haven't posted to this with the new app.

OS: Windows 8.1 Pro x64
BOINC Version: 7.6.33
GPU Driver: Radeon Crimson 17.3.3
CPU: AMD R7 1700 @ 3800MHz
GPU: AMD HD7970 @ 1250/1550

10 WU avg run-time - 32.08s

Looks like I haven't posted to this with the new app.

OS: Windows 8.1 Pro x64
BOINC Version: 7.6.33
GPU Driver: Radeon Crimson 17.3.3
CPU: AMD R7 1700 @ 3800MHz
GPU: AMD HD7970 @ 1250/1550

10 WU avg run-time - 32.08s

You have only one 32.08s and i can not find it... )))
your last 60 units are over 70s 80second ...
PLEASE do not "cherry pick" one short unit to benchmark
Several hundred or more can be take like some stable result.

Looks like I haven't posted to this with the new app.

OS: Windows 8.1 Pro x64
BOINC Version: 7.6.33
GPU Driver: Radeon Crimson 17.3.3
CPU: AMD R7 1700 @ 3800MHz
GPU: AMD HD7970 @ 1250/1550

10 WU avg run-time - 32.08s

You have only one 32.08s and i can not find it... )))
your last 60 units are over 70s 80second ...
PLEASE do not "cherry pick" one short unit to benchmark
Several hundred or more can be take like some stable result.

I am currently running 4 concurrently per card, because the efficiency is much better with all of the CPU time in the new tasks. Here are the pages with the units I ran singly to get a baseline:

https://milkyway.cs.rpi.edu/milkyway/results.php?userid=193178&offset=15360&show_names=0&state=4&appid=

https://milkyway.cs.rpi.edu/milkyway/results.php?userid=193178&offset=15380&show_names=0&state=4&appid=

No need to run 100s of units, I have ran 1000s of units at 4x, and the only invalids are due to other machines failing units. :)

Hey guys, I'm at around 57s with my 7970 Ghz Edition @ 1050/1500mhz running at the overclocked stock from the manufacturer and running 2 tasks at the same time. Running only one task was more like 60s.

It does look pretty low compared to all the other results here. I have tried running only with 35% CPU usage and it didn't change anything from the time compared to the 95% I was running before.

Is there maybe some magical setting I could have overlooked somewhere? I am running on a Ryzen 1800x and the most up to date stable Boinc, Windows 10 Pro and Crimson is updated as well.

Thanks for any input :)

Hi.
Here are some info and tips.

All my Tahiti XT (and one Tahiti LE) run 4 WU each at a time. I have standardized at 1GHz and 1000-1050mV GPU-voltage. I have stability tested each card for the lowest possible voltage running Furmark, Cinebench and M@h. Memory frequency does not influence WU time at all. My tip is to check what the actual GPU frequency is during heavy load. GPU-Z show only the value set. Use i.e. HWinfo (https://www.hwinfo.com/download.php) to show the actual running frequency.

My old computer 591552 with 4 Tahiti XT uses about 850Watts from the 230VAC grid. No water cooling, only one 140mm fan blowing into a standard midi-atx case with a Corsair RMi 1000W PSU. My Intel based PC's with two GPUs also run cosmology@home on the CPU. These machines consume about 470W each from the grid.

Regards Rally1965

Hi.
Here are some info and tips.

All my Tahiti XT (and one Tahiti LE) run 4 WU each at a time. I have standardized at 1GHz and 1000-1050mV GPU-voltage. I have stability tested each card for the lowest possible voltage running Furmark, Cinebench and M@h. Memory frequency does not influence WU time at all. My tip is to check what the actual GPU frequency is during heavy load. GPU-Z show only the value set. Use i.e. HWinfo (https://www.hwinfo.com/download.php) to show the actual running frequency.

My old computer 591552 with 4 Tahiti XT uses about 850Watts from the 230VAC grid. No water cooling, only one 140mm fan blowing into a standard midi-atx case with a Corsair RMi 1000W PSU. My Intel based PC's with two GPUs also run cosmology@home on the CPU. These machines consume about 470W each from the grid.

Regards Rally1965

Hey there Rally,

thanks a million for that input, I was totally unaware that GPU-Z is doing this!

I seem to have found the problem this way too. I only run at around 600 to 700mhz clock speeds while crunching MW@home, even when running 2 or 4 concurrent tasks. If I use my 7970s for mining, the full 1050mhz get used.

I already increased the power control setting to +20% in Crimson and that has helped a little bit, with the cards using 700 to 900mhz now (and a WU taking about 45 seconds now) but it still fluctuates quite a bit.

EDIT: It also seems that whenever I change a setting, the gpu uses the full 1050mhz for a dozen WUs or so and then throttles down again to 700-900mhz, increasing the time again.

*Sorry for the off-topic as well, but maybe it concerns someone else too that would like to benchmark ;)*

i do about 250 second for each unit. I've got a 1060 GTX 3GB, no OC'd.

AMD Phenom II X6 1100T running Windows 7 with an MSI RX 480 Armor 8gb OC GPU. Running one task at a time on the GPU with an average of about 110 seconds per task.

Like this in the task results:

Estimated AMD GPU GFLOP/s: 468 SP GFLOP/s, 94 DP FLOP/s
Warning: Bizarrely low flops (93). Defaulting to 100

---

6970 890Mhz/1375mhz no OC average: 94 sec per WU

GTX660 (2GB) 900 sec.
GTX1060(6GB) 275
GTX560 (1GB) 415
GTX650-Ti(1GB) 1115
GT720(1GB) 3000

GTX970 (4GB) 227 sec.
GTX650Ti (2GB) 1100

[url]Sorry for not replying in here for ages guys, been away from most forums lately, busy with other stuff.

Windows 7 Ultimate 64 bit - BOINC 7.6.33 - GPU Driver 14.1
Xeon E3 1270 V3 @ 3.7 GHz HT on
MW 1 task only - GPU @ 1065MHz - Memory @ 1555MHz 40 seconds - NO CPU tasks !
Same with CPU tasks limited to 88% - 43 seconds

dunx

P.S. Tried two tasks per GPU, and it messed up all four GPUs.... liking the sensible VRM temps though !

Err thanks for your time :), but what GPU is that?? lol ;)
And I assume you meant running 1 task at a time & not the time from just a single task?

Jozef J
No you don't need to run several hundred WUs to get a good average, lol, 5 WUs will do it as long as your machine is stable & follow the benchmark requirements.

Re your times, your post isn't clear as to which GPUs they belong to, I assume 39.xx s is from GPU1 & 41.xx is from GPU2 (why no decimal point? :p). But what is the 39s time for?
Your links no longer work naturally, I've left it too long, sorry :o.

Joses
Sorry, no idea what you mean by a cron job ;), but to give a reasonable accurate benchmark the CPU (& GPU) needs to be able to run MW @ nr 100%, so sorry I can't use your times. You could run just 1 core at nr 100% purely for benchmarking (BOINC preferences @50% for a dual core).

Tictoc
Thanks for your time :), will add it in.... awesome time btw! :D No1 by a long shot! I did a double take when I placed it, but when I saw your GPU clock I can see why, 1250 MHz is some o/c! :), water cooled?

DVDL
Is that time from at least 5 WUs?

Shu
Sounds to me like your card is running too hot & so throttling itself.

Darrell
Thanks for your time :), is that from an average of 5 or more WUs?
Re DP FLOPS, weird, according to my table a 480 should do 323 DP GFLOPS, was it throttling?

HassanShebli
Thanks for your time :), is that from an average of 5 or more WUs?

wb8ili
Thanks for all your times! :), are they from an average of 5 or more WUs?

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

Have added tictocs time to the MW AnandTech thread https://forums.anandtech.com/threads/milkyway-h-benchmark-thread-winter-2016-on-different-wu-gpu-cpu-times-wanted.2495905/#post-38654842 , will add other times there & here when I've had answers to the above questions.

---

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

Windows 7 Ultimate 64 bit - BOINC 7.6.33 - GPU Driver 14.1
Xeon E3 1270 V3 @ 3.7 GHz HT on
MW 1 task only - GPU @ 1065MHz - Memory @ 1555MHz 40 seconds - NO CPU tasks !
Same with CPU tasks limited to 88% - 43 seconds

dunx

P.S. Tried two tasks per GPU, and it messed up all four GPUs.... liking the sensible VRM temps though !

Err thanks for your time :), but what GPU is that?? lol ;)
And I assume you meant running 1 task at a time & not the time from just a single task?

dunx was running a pair of 7970s when those times were posted.

Tictoc
Thanks for your time :), will add it in.... awesome time btw! :D No1 by a long shot! I did a double take when I placed it, but when I saw your GPU clock I can see why, 1250 MHz is some o/c! :), water cooled?

GPUs are watercooled. I generally run them at 1200 MHz, because they hit the wall at 1200 MHz and it takes quite a bit of additional voltage to get them stable at 1250. Once that machine is back on Windows, I will have a new result to add. At least one of my 7970s will crunch MilkyWay at 1315ish.

Sorting out a few things, but I should have some Linux results, with the new AMD driver, on an R9 290 to add over the weekend. Making a push on Collatz at the moment, but I'll be back to MilkyWay in October. :)

Darrell
Thanks for your time :), is that from an average of 5 or more WUs?
Re DP FLOPS, weird, according to my table a 480 should do 323 DP GFLOPS, was it throttling?

Sorry, it is not a calculated average, but just an eyeball calculation from looking at the task results. An eyeball calculation of some of the new larger tasks is about 205 seconds.

The GPU is running fine, no throttling. Just a weird little message that shows up in each task result with the new GPU. It replaced an HD5850 back in March.

Task Results: http://milkyway.cs.rpi.edu/milkyway/results.php?userid=32442&offset=0&show_names=0&state=0&appid=

---

GTX 750 stock — 1000s without using browser and 1 free core
— 2000s without free core

---

DVDL
Is that time from at least 5 WUs?

Yes, i've done a ton of them. Not anymore, hope to get an AMD card sometime.

Ok folks, we have a minor problem, having been barely around I wasn't fully aware that the WU times from v1.42 (1.43?) to v1.46 had changed (slightly longer apparently), my bad!

Finally found the changelog here https://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=4125&postid=66313
(which wasn't put under a 'MW v1.46 release' title, thanks Jake! :p).

So all results post 1/5/17 will be put into a new table for v1.46.

For now I'm starting a new table in the AnandTech thread here https://forums.anandtech.com/threads/milkyway-h-benchmark-thread-winter-2016-on-different-wu-gpu-cpu-times-wanted.2495905/page-5#post-39153703

New benchmark requirement, 5+ result average to be from 227.2x credit WUs
(I don't think the .23 or .26 point difference will make much difference, but I will posts results from my rig to see for sure).

Darrell
Thanks for your time :), is that from an average of 5 or more WUs?
Re DP FLOPS, weird, according to my table a 480 should do 323 DP GFLOPS, was it throttling?

Sorry, it is not a calculated average, but just an eyeball calculation from looking at the task results. An eyeball calculation of some of the new larger tasks is about 205 seconds.

The GPU is running fine, no throttling. Just a weird little message that shows up in each task result with the new GPU. It replaced an HD5850 back in March.

Task Results: http://milkyway.cs.rpi.edu/milkyway/results.php?userid=32442&offset=0&show_names=0&state=0&appid=

Visual average is ok as long as the time varies little & you looked at at least 20 results (1 page). From 227.2x credit WUs I assume?
Btw, you didn't tell me what your GPU was o/c to, or what speed your CPU runs at.

GTX 750 stock — 1000s without using browser and 1 free core
— 2000s without free core

More info please, cpu type & speed? Is the 1000s from more than
And what do you mean by browser? As in web browser? That shouldn't make much difference anyway :).
Oh & new requirement, were the results you were looking at the 223.2x credit ones?

DVDL
Is that time from at least 5 WUs?

Yes, i've done a ton of them. Not anymore, hope to get an AMD card sometime.

Sorry, new question ;), were they all from 227.2x results?

---

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

Please use validated 227.23 credit WU results only, they must be from the MilkyWay@Home v1.46 app
Btw, watch out for the 227.26 credit WUs, they are very rare (approx. 1% of WUs atm), but despite their tiny increase in credit they take about 5% longer, at least on my HD 7970, ~56s vs 53s.

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 the above, 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.


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 v1.46 227.23 credit WU :-

HD 7970, GPU 1000 MHz (CPU, i7 4930k @4.1 GHz) ......................................... 53s ....... Assimilator1
RX 480 8GB, GPU o/c to? (CPU, Phenom II X6 1100T @?) ............................... 110s ........ Darrell

AnandTech MW benchmarking thread (with updateable table & req's) https://forums.anandtech.com/threads/milkyway-h-benchmark-thread-winter-2016-on-different-wu-gpu-cpu-times-wanted.2495905/

---

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