TR Forums

What do you guys know about BOINC (Berkley's version of Folding @ home) POEM @ home programs and how it is optimized for AMD HD 7000 series. I just bought an HD7950 after researching a little on this and it seems that the BOINC client is far better optimized for folding that the stamford version.

Here is a link to POEM@home via BOINC

http://boinc.fzk.de/poem/index.php?section=about

Let me know what you find.

Last I tried, (a few months ago) no one had any programs optimized for AMD Graphics core next (GCN); the new architecture for 79XX cards.
My AMD 7970 worked as well as a AMD 6850 on Folding.

Let me know if you find a BOINC project that is optimized.

Here is a thread I checked which convinced me. I just don't know how 480,000 PPD compares or translates to the Stanford Folding @ home PPD . The ratio must be 20 to 30 to 1.

http://www.amdusers.com/forum/showthrea ... de-to-POEM

Thanks for the link.

Looks like they are running multiple WU to improve utilization....but still only getting 55% usage on the card.

I must also note that my Radeon HD 7950 with 28 CUs is running 1200MHz core and 1700MHz memory (4000 GigaFLOPS SP), whilst my old Radeon HD 6850 with 12 CUs is only running at 860Mhz core an 1250MHz memory (1700 GigaFLOPS SP). Running 5 work units with my 7950, each taking 45 minutes to complete, equals 9 minutes per completed WU at 55% utilization. Meanwhile, my 6850s running 5 work units take around 77 minutes to complete, so 15.5 minutes per completed task with 90% utilization.

Running many WU leads to CPU/bus bottleneck...

Sounds like he is certainly improving on the standard performance, but still using a kludge instead of natively optimized client.

I'm not up for that much work right now.
But again, thanks for passing on the information.

Well......I installed an HD7950 and began FOlding with BOINC Poem @ Home and it has "scored" 1.7 Million Points in a week and a half! Does anyone know how much "work" this is as compared to Folding @ HOME? In other words what is the ratio of POEM @ Home PPD to Folding @ Home PPD.... I am averaging about 80,000 Points Per unit but it seems like approximately 200,000 PPD. What would this represent in terms of work accomplished vs Stamfords FOlding @ Home where due to optimization issues an HD7950 may not even crack 10,000 PPD.

Boincs POEM @ home is optimized for Radeon 7000 GCN architrecture from what I have read.

Since it is a completely different project running different code and different simulations, I doubt there is a way to meaningfully compare the two.