Posts by Stef

log in
1) Message boards : Technical News : GW follow-up run #3 (S6BucketFU3UB) (Message 143705)
Posted 17 Sep 2015 by Profile Stef

Why don´t I get new tasks, even when they are obviously available. ?????

Same for me. There are numerous "task to send" but I don't get any, although my buffer for CPU-tasks is only about 10h.
2) Message boards : Cruncher's Corner : HD 5870 worth running... (Message 143555)
Posted 15 Sep 2015 by Profile Stef
It's already 5 years ago, when AMD bought ATI. Guess there aren't many ATI cards around anymore.
3) Message boards : Cruncher's Corner : Low memory clock on Maxwell2 cards (960/970/980, probably Titan X) (Message 143382)
Posted 12 Sep 2015 by Profile Stef
On linux i use nvidia-smi to set it to the maximum frequency.
For my card it is "nvidia-smi -ac 3600,1544". first is the memory clock (half the DDR data rate frequency), second the gpu clock.
To show the supported clock combinations run "nvidia-smi -q -d SUPPORTED_CLOCKS"
For your card I assume "nvidia-smi -ac 3505,1531"
4) Message boards : Science : Save the dates ... (Message 143282)
Posted 10 Sep 2015 by Profile Stef
Science run does not mean data analysis by Einstein@Home.

Ah I got that wrong, thank you.
5) Message boards : Science : Save the dates ... (Message 143243)
Posted 9 Sep 2015 by Profile Stef
Advanced LIGO first science run "O1" start Sep 14.

Will it make use of AVX? (ref.)
6) Message boards : Problems and Bug Reports : Problem running more than 2 GPU WUs at a time (Message 143102)
Posted 7 Sep 2015 by Profile Stef
That did the trick, thank you.
Should have been more patient :)
7) Message boards : Problems and Bug Reports : Problem running more than 2 GPU WUs at a time (Message 143100)
Posted 7 Sep 2015 by Profile Stef
Hi,

I've tried to run 3 GPU WUs at a time, but i only get two at a time to work. What I did:

- Set all GPU utilization factor to 0.33 for "work"
- Added the host to "work"
- Clicked update in the boingmanager

The log says it is added to "work":
Mon 07 Sep 2015 11:40:13 AM CEST | Einstein@Home | Computer location: work

But there are still only running two GPU tasks (each 0.2 CPUs + 0.5 NVIDIA GPUs)

What have i done wrong?
8) Message boards : Cruncher's Corner : Pushing too hard (Message 143072)
Posted 6 Sep 2015 by Profile Stef
I think S6Bucket #2 needs to be finished before the start of #3, that's why the short deadlines.
9) Message boards : Cruncher's Corner : Problem with a R9 390X when run 2 or more WUs at a time. (Message 142923)
Posted 31 Aug 2015 by Profile Stef
I didn't find any newer high-end AMD GPUs in the top 100 hosts list, so it definitely is a problem.
10) Message boards : Cruncher's Corner : AMD Fury GPU (Message 142827)
Posted 26 Aug 2015 by Profile Stef
Yes, would be nice if a developer could have an eye on that.
I'm planning to buy a new card but with this problems on recent high-end/enthusiast AMD cards, it's difficult to decide. Especially with the upcoming R9 Nano.
11) Message boards : Cruncher's Corner : Best single GPU rig and AMD vs. NVIDIA (Message 142784)
Posted 24 Aug 2015 by Profile Stef
My board only has 1 PCI-E slot.

But I'm a bit uncertain now regarding the AMD R9 cards, since in some cases they only seem to run one WU at a time, without producing invalid results.
12) Message boards : Cruncher's Corner : Best single GPU rig and AMD vs. NVIDIA (Message 142756)
Posted 21 Aug 2015 by Profile Stef
I think my PSU has a good efficiency (Bequiet E9 400W). I'll wait until the announcement of the R9 Nano GPU next week.
But relating to the R9 Fury thread (http://einstein.phys.uwm.edu/forum_thread.php?id=11352) it may cause some trouble. I haven't seen any of the HBM GPUs in the top 100 list.
13) Message boards : Cruncher's Corner : Best single GPU rig and AMD vs. NVIDIA (Message 142745)
Posted 21 Aug 2015 by Profile Stef
I don't mind energy consumption, since i only run it during winter and it is assisting the heating, so I need the exact amount of energy less for heating.

I was looking in the top hosts statistics, because the RAC is including the possibility of multiple instances of e@h running, so I don't have to look into the details.


Let's say I'm willing to spend 200€, I have the choice between a GTX 960 and R9 380.

So If I stick to e@h, AMD is the way to go?
14) Message boards : Cruncher's Corner : Best single GPU rig and AMD vs. NVIDIA (Message 142734)
Posted 20 Aug 2015 by Profile Stef
I was trying to find the fastest single GPU rigs, but i didn't found a more efficient way than scrolling through the top hosts list and look for the single GPU ones.
Is there a more efficient way to do this?

And the second question regarding the latest CUDA enhancements:
With which card(s) do I get the most credit per $?
15) Message boards : Cruncher's Corner : AMD Fury GPU (Message 142215)
Posted 24 Jul 2015 by Profile Stef
I think those high-end GPUs need a lot of units in parallel to benefit from their massive amount of ALUs, regardless what the GPU usage says.

Would be nice if you could compare them with 4x to 8x.
16) Message boards : Cruncher's Corner : GPU computing benchmark site tip? (Message 141664)
Posted 26 Jun 2015 by Profile Stef
Computerbase (german) did some computing Benchmarks; Musemage, CompubenchCL, Computemark, Folding@home, GPUPI and luxmark:

http://www.computerbase.de/2015-06/amd-radeon-r9-fury-x-test/8/
17) Message boards : Cruncher's Corner : GPU computing benchmark site tip? (Message 141648)
Posted 25 Jun 2015 by Profile Stef
It depends very much on the application. Application A can run faster on one card and application B on the other.
The new R9 Fury X should be very fast with double precision calculations, at least compared to the NVidia.

I'm very curious how it performs on einstein (it doesn't use double precision though).
18) Message boards : Problems and Bug Reports : BRP4 Intel GPU app feedback thread (Message 141443)
Posted 10 Jun 2015 by Profile Stef
I did a "echo 0 > /sys/module/i915/parameters/enable_cmd_parser" and patched the Kernel (4.0.4) with https://01.org/zh/beignet/downloads/linux-kernel-patch-hsw-support
But I only ran one WU, which turned out invalid.
19) Message boards : Problems and Bug Reports : BRP4 Intel GPU app feedback thread (Message 141434)
Posted 10 Jun 2015 by Profile Stef
I compiled beignet from the latest git. It compiled and worked, but it did not verify against BRP6-cuda32-nv301.
So the latest version of beignet known to work with einstein is 0.9.2.
20) Message boards : Cruncher's Corner : Times (Elapsed / CPU) for BRP5/6/6-Beta on various CPU/GPU combos - DISCUSSION Thread (Message 140865)
Posted 6 May 2015 by Profile Stef
Well, luckily I just sold the 750 Ti for exactly the same price I bought it 10 months ago - because of currency loss. So all I paid was the energy, which I would have to pay anyway to heat the house.
But I just found out I'm in the wrong thread. So I'll start a new one in september/october :)


Next 20

Home · Your account · Message boards

This material is based upon work supported by the National Science Foundation (NSF) under Grants PHY-1104902, PHY-1104617 and PHY-1105572 and by the Max Planck Gesellschaft (MPG). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the investigators and do not necessarily reflect the views of the NSF or the MPG.

Copyright © 2016 Bruce Allen