|
41)
Message boards :
Number crunching :
Double your task throughput on Linux
(Message 5784)
Posted 20 Aug 2022 by  Tom M
Post:
I was apparently tired yesterday. Meant to toggle the WCG downloads off but managed to do it to the U@H downloads instead. So wasn't processing any U@H this morning. There will be a blip in my production.... Tom M |
|
42)
Message boards :
Number crunching :
Milestones
(Message 5783)
Posted 20 Aug 2022 by Tom M
Post:
I just looked at your computers. It looks reasonable that the 2 machines you have that are running at 1M+ will be joined by 2 more (same brand/model/operating systems). Plus you rejoice in a "full on" farm of 18 systems registered at U@H. I remember some of those older server CPU model numbers. Back when I was making more money I had a couple of dual CPU motherboards using them. Salute! Tom M |
|
43)
Message boards :
Number crunching :
Milestones
(Message 5781)
Posted 20 Aug 2022 by Tom M
Post: Hello... Tom... are you still there? You need to breathe now! So are you PROUD of being humble? ROFLing.... Thank you for the breathing advice. I will start panting now..... |
|
44)
Message boards :
Number crunching :
Milestones
(Message 5780)
Posted 20 Aug 2022 by Tom M
Post: Hope to get a bit higher with my Epyc 7601 than 500k but who knows? And my EPYC system has made it past 600,000 on its U@H processing. Who knows how far it will get? |
|
45)
Message boards :
Number crunching :
Double your task throughput on Linux
(Message 5779)
Posted 20 Aug 2022 by Tom M
Post: I may have got the NBIO section wrong. But definitely in the SMU Options whatever upper level structure is in front of that. That sounds like a good plan. Where are my hands? I need to sit on them. :) The EPYC system has been running since yesterday. Hasn't crashed (yet). WCG has started trickling out. I have it limited to 2 CPU tasks. Just wondering how high the U@H RAC will go above 600,000 (its above it now). Tom M |
|
46)
Message boards :
Number crunching :
Double your task throughput on Linux
(Message 5777)
Posted 19 Aug 2022 by Tom M
Post: Should be in the same place as the cTDP setting. Apparently the 7601 doesn't support CBS>>NBIO>>SMU options. As far as I can tell the MB has the latest Bios version 2.40 However, so far, it is running with a cTDP at 200 watts and the CPU speed is up and is now claiming to be running just short of 2.7 GHz the "all core" limit. Tom M |
|
47)
Message boards :
Number crunching :
Double your task throughput on Linux
(Message 5774)
Posted 19 Aug 2022 by Tom M
Post: [quote] Running on Power determinism and both cTDP and PPT set to the max 200W allowed makes all the cores lock to the max turbo speed of 3.35Ghz when loaded with Universe tasks. /quote] I need to see if I can locate a "PPT" setting for the 7601 cpu. It sounds like that would be an excellent experiment after I get done with the 90% load experiment. Early results sound encouraging. Within a day of balancing the CPU to 90% load (which meant setting the project max file in the U@H folder to 53 threads/cpus) the RAC passed the 600,000 mark. Tom M |
|
48)
Message boards :
Number crunching :
Double your task throughput on Linux
(Message 5772)
Posted 18 Aug 2022 by Tom M
Post:
I had to drop the CPU processing count in u@h down to 40 (from 53 for the 90% load) before the reported GHz on the CPU got up to near the 2.7 GHz "all core" speed the specifications for the Epyc 7601 claim. Somehow I don't think that would produce a net "gain" in total production :) Tom M |
|
49)
Message boards :
Number crunching :
Double your task throughput on Linux
(Message 5771)
Posted 17 Aug 2022 by Tom M
Post:
It certainly looks like no change in production. In other news It looks like I can reduce the ratio of wall clock time to CPU processing time to under 1 minute by requiring a full thread per GPU task. To be more confident of this finding I have switched the GPU processing back to 0.5 CPU thread to GPU thread. Both CPU load indicators show a slight increase in the CPU load (from 95-96% to 97-98%). Actual change in production as shown by GUI graph/RAC scores has been nearly flat, so far. This should give me an excellent way to see if there is a result when I drop the CPU processing load to 90%. My impression is you get more CPU production at 90% load then you get at 96-99% CPU load. This is looking like I will be able to validate that impression one way or the other. Tom M |
|
50)
Message boards :
Number crunching :
Double your task throughput on Linux
(Message 5765)
Posted 15 Aug 2022 by Tom M
Post:
First results are not encouraging. The production graph/RAC are showing a slight drop. And some of the differences between wall clock and CPU time have increased to near 4 minutes from 2.5 minutes. But still need a longer period of experiment to confirm the trend. On theoretical grounds I have been running with the Virtual CPU setting off in the bios. Just turned that on. Curious to see if any production changes. Check back in, in the morning. Respectfully, Tom M |
|
51)
Message boards :
Number crunching :
Double your task throughput on Linux
(Message 5762)
Posted 14 Aug 2022 by Tom M
Post:
Since it appears production went down overnight when I set the cTDP to the nominal 180 watts. I have reset that to "auto". I just set the U@H project limit in its app_config.xml to 64. It appears it is now running at 60 (I think I counted) because I also set the gpus to reserve a half a thread per gpu task (4 gpu tasks). As predicted the CPU load went to 99% under the Psensor. And the System Monitor is showing 62 cpus running at over 95%. I am still running two CPU threads "idle" so they are getting as low as 87%. Should have some results by the end of the day. Tom M |
|
52)
Message boards :
Number crunching :
Double your task throughput on Linux
(Message 5757)
Posted 14 Aug 2022 by Tom M
Post:
On the EPYC system. I am currently running 96.87% of 64 cores (~62) on the BOINC Manager. I am currently running 4 gpu threads, 1 primegrid thread, 2 TN-Grid threads and have a project limit of 50 on the U@H threads. It currently looks like the wall clock time is about 2.5 minutes slower than the CPU time. According to Psensor I am running @90% CPU load. System Monitor is also showing threads running below 50% on occasion. I can drive the CPU load up to 99% but haven't tried that when I was paying attention to the wall clock vs. CPU time. Sounds like another experiment is coming. I have just set the cTDP from "auto" to 180 which is the "nominal" maximum cTDP. I have also turned off the Virtual Machine setting in bios. The determinism slider is still on "power". Once the Psensor for how hot the CPU is went back up to 81C the CPU slowed back down to under 2.6 GHz instead of slightly over it. Tom M |
|
53)
Message boards :
Number crunching :
Double your task throughput on Linux
(Message 5754)
Posted 13 Aug 2022 by Tom M
Post: Context. In a system running both CPU and GPU tasks. What ratio (percentage too) are you running each of your systems at in terms of available threads/cpu's vs. Boinc Manager thread Limit? For instance a 3950x has 32 threads aka: cpus on the Boinc Manager. If you are running 30/32 = 93.75% And what is your maximum # of U@H project threads set at in your app_config.xml file? For instance "32" What I am looking for is additional examples of where you are finding your maximum production point in U@H. Thank you. Tom M |
|
54)
Message boards :
Number crunching :
Milestones
(Message 5749)
Posted 12 Aug 2022 by Tom M
Post: Hope to get a bit higher with my Epyc 7601 than 500k but who knows? It is getting closer and closer to 600K. <holding breath> <turning blue> <gasp> |
|
55)
Message boards :
Number crunching :
Milestones
(Message 5735)
Posted 6 Aug 2022 by Tom M
Post: I've been able to keep all my PC hosts above 500K RAC so far. Even while also doing TN-Grid and yoyo CPU tasks. Congratulations. Hope to get a bit higher with my Epyc 7601 than 500k but who knows? Tom M |
|
56)
Message boards :
Number crunching :
Milestones
(Message 5733)
Posted 5 Aug 2022 by Tom M
Post: Just passed a half million on this machine. https://universeathome.pl/universe/show_host_detail.php?hostid=628118 |
|
57)
Message boards :
Number crunching :
Milestones
(Message 5732)
Posted 5 Aug 2022 by Tom M
Post: One of my favorite threads on some other projects is "Milestones". I expect to have another system passing the half-million mark for its U@H RAC very soon. And I wanted a place to brag about it here. The commonest "brag" however is the total RAC a system/user has run up. Something very impressive about "1 BILLION" or "1 TRILLION" totals.... It begins to feel at least Galaxy sized... Tom M |
|
58)
Message boards :
Number crunching :
Best Bang for the Buck?
(Message 5697)
Posted 4 Jul 2022 by Tom M
Post:
After examining my capital budget and my sensitivity to cost I bought the cheapest/new Epyc 7601 I could find. Since it is from China it is going to take a while to get here. I have tested nearly all the ribbon cables I currently own. None are working. So it LOOKS like I am going with a 7601 (32c/64t) + M2. SSD + 2 RTX 3080 ti FE gpus for now (as soon as they arrive). AFAICT, it looks like the 7601 will be a straight linear scale up of the 7251 (8c/16t) that I have. The specs for the multi-core boost appear to be identical to what I am getting now. Tom M |
|
59)
Message boards :
Number crunching :
Best Bang for the Buck?
(Message 5696)
Posted 3 Jul 2022 by Tom M
Post:
I would consider them except they are not ROME cpus. The increase in efficiency/processing speed based on the reviews is significant. If I limit myself to new Rome cpus in the USA (faster shipping) there are 7302's (16c/32t) for under $500. If I add refurbished I can find a 32c/64t QS for about $750 in the USA. And a 64c/128t QS with a missing memory channel (if the stars align) for under $1,300. If I shop for the cheapest 7601 (overseas currently) I can get a lower than lowest Rome price but loose the generational gains from Naples to Rome. I am going to chase the refund for my last attempted EPYC CPU upgrade this next week. And continue testing Ribbon Cables on the gpus. If I was content with 2 GPUs on the MB I could stop testing/buying ribbon cables. Tom M |
|
60)
Message boards :
Cafe :
Humor including Jokes, Puns, Limericks, Shaggy Dog stories etc.
(Message 5695)
Posted 3 Jul 2022 by Tom M
Post:
Ouch! |
(English)
(Polski)