PCIe Bandwidth, important for Distributed Computing?

[TennesseeTony]

I know for fast action, graphics intensive gaming, maximum bandwidth is required. But I also know for things like BitCoin mining, only having 1x going to the GPU is perfect, overkill perhaps.

What about for the various projects of BOINC, and Folding at Home? Is 1x adequate?

My only experience was with multiple R9-280X's, running POEM and Milkyway. I did not see any decrease in output using 1x on Milkyway, but due to the projects task limit, time limit between refreshes, and the 21 seconds per task, I was almost always out of work, so I'm not real sure.

Likewise, POEM seemed to be unaffected by 1x bandwidth, but the system was a bit unstable by that time, there were card issues and possible PSU issues, so again I'm not 100% sure there was not a reduction in task performance.

Anyone have a definitive answer on which projects need more than 1x? Personal experiences?

 

[[H]Coleslaw]

We had a user start collecting some data in regards to Folding here: https://hardforum.com/threads/pcie-speed-folding-performance.1902126/#post-1042373671

 

[TennesseeTony]

Much appreciated Coleslaw! Perhaps I can add some 1x data, as they only tested down to 8x.

 

[TennesseeTony]

Coleslaw, if you wouldn't mind linking/reposting this to that thread, perhaps it can be of some use to someone.

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

I have quickly downloaded FAH Bench and ran the GUI, and ran 180 second tests, one at PCIe 2.0 16x and the other at 1x using this contraption (link).

My test bed is a BIOSTAR H81S2, with a 3.0GHz Pentium dual core (G3220, 22nm). The GPU is a lowly Radeon R9-280X (HD7970).

At 1x, the FAHBench software reported these results on Tahiti, Open-CL, accuracy check enabled, 180 seconds, dhfr task, single precision, NaN check disabled: 38.5408, 23558 atoms.

At 16x/2.0: 42.3566, 23558 atoms.

1x was only 91% of 16x, losing 9% of it's potential. The test was not duplicated, and this GPU hardly stresses the PCIe bus the way a more modern card can. Still a substantial difference though.

I will likely try a 980Ti in the coming days. That should be far more interesting, but for now I've got to get off my butt and get some things done.

Edit #1: And I'm impatient, 980Ti results coming soon...

Edit #2: GTX 980Ti = 81.4012 at 16x/2.0; at 1x it was, wow, 41.777.

Huge hit, only 51.3% of the 980Ti's potential was extracted using a 1x interface. I did notice that MSI afterburner was reporting BUS usage at 12-13% at 16x, but only 6-7% at 1x. That seems to scale with the results from FAHBench.

10esseeTony.

 

[[H]Coleslaw]

Done

 

[TennesseeTony]

Thank you Sir!

 

[kowalabearhugs]

Thank you, Tony