Posts by Shaggie76

1) Message boards : Number crunching : Best bang -for-buck GPU 2018? (Message 1928017)
Posted 17 days ago by Profile Shaggie76
Post:
You can start with Shaggie76's often updated chart of GPU FLOPS: Theory vs Reality

I wish there were new cards released to motivate me to run another scan.
2) Message boards : Number crunching : The best 'off the shelf' laptop for seti@home (Message 1924721)
Posted 15 Mar 2018 by Profile Shaggie76
Post:
From what I can tell modern CPUs hit a thermal limit and throttle the clock-rate which is especially bad for laptops -- short sprints are fine but endurance runs get too hot and the chip slows down. My Kaby Lake quad laptop does 20-minute builds at almost exactly the same speed as a 4 year old dual laptop and diagnostics indicate that the chip has down-clocked to cope with the heat.
3) Message boards : Number crunching : SETI@home v8.22 Windows GPU applications support thread (Message 1918372)
Posted 11 Feb 2018 by Profile Shaggie76
Post:
Seemed like we had a winner but then lockup at 95.38%. After restart it resumed from a CP from around 91% and locked up again shortly afterward.

I'm growing uncomfortable with so many hard shutdowns on this system -- I'm over 10 now and I'm worried about FS corruption on a machine I need to work on. I've run it 5 or 6 times in the benchmark but in BOINC it always brings down my machine.

If you aren't really interested in anything diagnostic I'll abort this task and see if any others cause problems.
4) Message boards : Number crunching : SETI@home v8.22 Windows GPU applications support thread (Message 1918364)
Posted 11 Feb 2018 by Profile Shaggie76
Post:
Ok, It wasn't in the BOINC or the project directory but ProgramData\BOINC\slots\0\state.sah

I will kill that and try again in a minute. I should also note that I've tried running the task 3x back to back in the bench by duplicating the WU and no problems there.
5) Message boards : Number crunching : SETI@home v8.22 Windows GPU applications support thread (Message 1918278)
Posted 11 Feb 2018 by Profile Shaggie76
Post:
simple deletion of state.sah in slot should do.


I couldn't find that but I removed
<active_task_set>
<active_task>
...

from client_state
I also deleted the partial result file from the app directory. I will try again.

If this doesn't work the other thing I can try is replicating the use_sleep option to the offline benchmark.
6) Message boards : Number crunching : SETI@home v8.22 Windows GPU applications support thread (Message 1918206)
Posted 11 Feb 2018 by Profile Shaggie76
Post:
It's a laptop; I won't bother with eGPU otherwise.

Any other system you can run the eGPU on?
If it runs OK on all other systems & GPUs, then that just leaves Thunderbolt as the issue.
Be it drivers or hardware.
Any updated Thunderbolt drivers/BIOS for the motherboard or the eGPU unit about?


Well I have 3 matching cards in another machine -- I'm not really interested in moving hardware around because, like I said, the card runs 3D apps just fine -- it's been running our game for 4 hours now, it's cleared over a hundred SETI work units, and the same workunit runs start to finish in the benchmark.

And yes, I checked drivers.

I hacked the XML to clear the checkpoint but still locked up after-- I must not have done it right. I'm not sure if there's a proper way to do that.
7) Message boards : Number crunching : SETI@home v8.22 Windows GPU applications support thread (Message 1918171)
Posted 11 Feb 2018 by Profile Shaggie76
Post:
Full disclosure: this is a 1080 Ti running in an eGPU enclosure over Thunderbolt 3

So there's a whole other kettle of fish thrown in to the mix.
Does it lock if the Video card is physically in that machine? (ie take the Thunderbolt connection out of the mix. I've seen several reviews where certain benchmarks won't complete on certain systems when running GPUs in external enclosures over Thunderbolt, yet will run on a different motherboard. Various hardware/BIOS/driver incompatibilities between different motherboards & external units).


It's a laptop; I won't bother with eGPU otherwise.

-g
8) Message boards : Number crunching : SETI@home v8.22 Windows GPU applications support thread (Message 1918149)
Posted 10 Feb 2018 by Profile Shaggie76
Post:
Put that task into offline bench and re-run - will it lock?


I set it up and ran it through once -- no lockup -- but in doing so I noticed the checkpoint file and wondered if that might be to blame. Can I bench 'from the checkpoint somehow? I'm not sure how to do this in the bench but I'll fiddle around a bit.
9) Message boards : Number crunching : SETI@home v8.22 Windows GPU applications support thread (Message 1918147)
Posted 10 Feb 2018 by Profile Shaggie76
Post:
Shaggie, do you run Process Lasso as well on your machines?


I do not.
10) Message boards : Number crunching : SETI@home v8.22 Windows GPU applications support thread (Message 1918113)
Posted 10 Feb 2018 by Profile Shaggie76
Post:
I noticed the thread here about the latest NVidia drivers changing some OpenCL stuff so I went back to 390.65 and let it try again - locked up again.
11) Message boards : Number crunching : SETI@home v8.22 Windows GPU applications support thread (Message 1918061)
Posted 10 Feb 2018 by Profile Shaggie76
Post:
In short, https://setiathome.berkeley.edu/result.php?resultid=6388966004 and possibly other tasks hard-lock my machine.

Yes, a hard-lock is not your fault, but in my game dev experience often when a player's machine locks up or blue screens it's a sign that we're abusing the driver in a way that it can't handle and we usually make an effort to play nicer. I hope you have a similarly positive attitude.

I would also offer the same infuriating anecdotal evidence our players give me: I can do other heavy-lifting with the card for hours and it doesn't lock up.

Full disclosure: this is a 1080 Ti running in an eGPU enclosure over Thunderbolt 3 (Razer Core v2). It crunched for a few days just fine but cannot get past this one task. It does no CPU tasks and does not crunch while I'm using the machine -- I'd hope that the risk of other things using the CPU causing timing issues would be minimal. I've whittled it down from 2 tasks at once to 1 task with a reduced cmdline: <cmdline>-use_sleep -tt 10</cmdline>. I have currently paused BOINC on this machine so that the task doesn't expire.

There's nothing in the event log that has a clue. I'm on the latest NVidia drivers. This machine is pretty new -- only a few weeks old now -- and has crunched about 34K credit so far -- maybe 24 hours of crunching distributed over the last week or so.

If there's any diagnostics you'd like me to run just let me know how I can be of assistance. I have VS2015 but no OpenCL libraries or anything -- I'm not sure if there's a debug shim like there is for DirectX that I can run with but if you give me directions I'm happy to try.

-g
12) Message boards : Number crunching : GPU FLOPS: Theory vs Reality (Message 1909892)
Posted 31 Dec 2017 by Profile Shaggie76
Post:
I wonder how badly you think mismatched cards in a system affect your numbers.
For instance my 4x1070 computer is actually a 1070/980/1060/1060.

The script omits all multi-GPU systems for exactly this reason.
13) Message boards : Number crunching : GPU FLOPS: Theory vs Reality (Message 1909711)
Posted 31 Dec 2017 by Profile Shaggie76
Post:
Average of Median 60% Credit/Hour from my last run:

NVIDIA GeForce GTX 1070: 908.8006478
NVIDIA GeForce GTX 1070 Ti: 941.2174884

So approximately 3.6% more credit from a wide sampling of computers and tasks.

From your data below, average power:

NVIDIA GeForce GTX 1070: 111.5 W
NVIDIA GeForce GTX 1070 Ti: 117.7 W

So approximately 5.6% more power in this specific case.

I noticed your 1070's were rated for 166W not the stock 150W -- so maybe clocked a bit higher and drawing a bit more power too?
14) Message boards : Number crunching : GPU FLOPS: Theory vs Reality (Message 1909613)
Posted 30 Dec 2017 by Profile Shaggie76
Post:
New scan shows the 1070 Ti's doing a tiny bit more work than regular 1070's but probably use more power than it's worth.

15) Message boards : Number crunching : GPU FLOPS: Theory vs Reality (Message 1905934)
Posted 9 Dec 2017 by Profile Shaggie76
Post:
After all the server problems last week I didn't want to put strain on them running my script to crawl through the database. Maybe if they stay up this week I'll try again.
16) Message boards : Number crunching : GPU FLOPS: Theory vs Reality (Message 1903225)
Posted 28 Nov 2017 by Profile Shaggie76
Post:
I ran another scan today and there still aren't enough 1070 Ti's in circulation yet.
17) Message boards : Number crunching : GPU FLOPS: Theory vs Reality (Message 1900619)
Posted 12 Nov 2017 by Profile Shaggie76
Post:
I ran a few scans over the last two weeks and combined the data together to give a picture of some rarer cards that might not otherwise have enough valid results to get a picture from. There aren't enough Pascal Titans yet but this scan has some RX Vega parts in mix (sadly it does not seem well optimized for SETI by default). There aren't enough 1070 Ti hosts yet either (2 by my count).

I took a stab at hooking up data for Intel IGPs -- the TDP values are probably wrong so don't pay too much attention to the CPWH chart for them -- I was mostly curious about throughput and it's pretty abysmal as you'd expect. I probably won't include those again since the CPWH is misleading.

To try to keep the charts manageable I've started omitting some of the older generation cards; I assume that since you're probably looking at this data to help guide setting up machines it's unlikely that you'll be worried about vintage parts but I could be wrong.

18) Message boards : Number crunching : GPU FLOPS: Theory vs Reality (Message 1899570)
Posted 6 Nov 2017 by Profile Shaggie76
Post:
Still not interested, sorry.
19) Message boards : Number crunching : GPU FLOPS: Theory vs Reality (Message 1899198)
Posted 4 Nov 2017 by Profile Shaggie76
Post:
I don't know for sure but I suspect that mining tasks use integer arithmetic but SETI@home uses floating-point math; AMD cards might offer better integer performance per watt which could explain the miners' enthusiasm.
20) Message boards : Number crunching : GPU Advice 980 vs RX 480 vs 1070 (Message 1898682)
Posted 2 Nov 2017 by Profile Shaggie76
Post:
Me latest benchmarks might help you decide -- the RX 480 has been disappointing since it shipped -- I'd hoped the performance would improve but it hasn't really.

I'm keen to measure but I'm guessing the 1070 Ti won't perform much better than the 1070 because the memory bandwidth per core is lower -- even the 1080 isn't that much faster than a stock 1070.


Next 20


 
©2018 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.