Pretty old thread to bring back up but I was watching the video again in the OP.
I didn't re-read the whole thread again, but don't think anyone caught how he compaired
how much faster core 17 is suppose to be from core 15 and 16. If this had been done on even ground I don't think we would see much of a difference.
They used PPD of the old and PPD of the new to make core 17 look like its faster lol
where as "
if" core 15 and 16 had had QRB as core 17 does I don't think we see that much of a difference.
So in reality core 17 is at most no faster than what we had it only allows more PPD because of QRB.
I don't think the work is getting done any faster now than before, actually some of the newer WU take longer to complete.
And core 11 never got any love to start with it still has the low base points it always had were core 15 did get some boost
or change to it points value along the way.
Edit: not to mention core 17 needs use of a full CPU core in order for it to do its work at max speed where core 15
and core 11 only need a very small use of a core 1 to 2 % of a core not 100% of a single core for Nivida video cards.
So this in turn hurt how much PPD you could make from SMP folding and its QRB.
Maybe one day Nvidia will fix the core usage but for now we are taking a big hit on SMP folding because of core 17 were we didn't before.
So you also need to remove some of the PPD so called gained by core 17 to compensate for the loss to SMP folding PPD.
post edited by TheWolf - Monday, November 25, 2013 10:29 AM