All rigs to stock..... (?)

[orion456]

Due to a medical situation I will be off work for an extended period of time and unable to tend my gaggle of beasts. I expect they will either slowly shut down or suddenly go off line as events conspire against them. In order to reduce problems and hopefully extend service, I have stopped overclocking and set all rigs to stock frequency and voltage. I have not yet set all the GPUs to stock but they are very mildly overclocked if at all.

I expected an immediate point drop because taking a quad from 3.4ghz go 2.4is a 40% loss of processing power. BUT to my surprise FahMon reports my totals have increased with no loss of SMP processing results....how could that be? The drop in heat production is significant, so if stock produces the same results as OC; what have I been doing all this time?

I assume its related to the GPU/SMP interactions in some way. Or is it that a quad is starved for data at 3.4ghz and the FSB really needs to be increased to make a sufficient supply?

[ChasR]

My guess would be you had some p2668s and p2669s replace p2605s on the SMP instances.

[orion456]

My guess would be you had some p2668s and p2669s replace p2605s on the SMP instances.

All the SMPs are running 2665s which I believe they are running before except one that had a 1760 pointer. Meantime my GPUs have jumped to the point where one making 5800ppd is now doing 6380ppd on a 480 points WU; seems strange.

[LandShark]

All the SMPs are running 2665s which I believe they are running before except one that had a 1760 pointer. Meantime my GPUs have jumped to the point where one making 5800ppd is now doing 6380ppd on a 480 points WU; seems strange.

the GPU ppd sounds like it's due to the recent core update to 1.13/1.15v if you haven't update it in the past week or two.

the SMP ppd, well, depends on your FahMon setting, if it's set to calculate all frame, then it will take a quite to average out again. if it's set to last 3 frame, you'll get a much more up to date ppd.

p.s. btw, hope you are doing ok!!

[orion456]

the GPU ppd sounds like it's due to the recent core update to 1.13/1.15v if you haven't update it in the past week or two.


My stats for the last five days:

10.19.08 50,528 83
10.18.08 47,722 88
10.17.08 42,824 79
10.16.08 43,967 73
10.15.08 37,588 67

Two days ago I set all my rigs to stock CPU settings and for the last two days my point total has gone up +5000 ppd. Looking at other's points, I don't see any significant up trend for the group as a whole; as you would expect if it as a core change or some system wide effect. In my case, I would have expected a substantial drop especially going from 3.4ghz to 2.4; something's strange. :screwy: It might just be statistics, but if so, I am an anomaly!

I also noticed that it makes almost no difference if you have two cores working on a problem or three. I suppose this is the same reason two sets of two cores are better than one set of four cores.

ps: surgery on the 28th, if I report, you will know I'm still among the folders.

[StaTek]

Those stats are weird...and not what I would expect either.

Hope your surgery goes well & you have a speedy recovery!

[harlam357]

Hope your surgery goes well & you have a speedy recovery!

Double that... here's to your speedy recovery. :beer:

My guess is that your GPUs have updated their core to v1.15 and you're seeing a nice spike in PPD because of it.

Believe me, the CPU OCs do make a difference... just not enough to offset the increased GPU PPD.