GekkoScience Compac Overclock Contest

[kilo17]

I have decided to start a new thread here about an Overclock contest that is being discussed on the Compac Thread in the Hardware section:

We are looking at holding the contest this Friday and it will last 24 hours - Times to be announced.

I will post details as we come to an agreement about the rules and entry fee if there is one.

[kilo17]

Reserved for entry names of the contest

[cavaliersrus]

im not entering but i really want to follow how people are going to cool there sticks in the competition

[Biodom]

I have decided to start a new thread here about an Overclock contest that is being discussed on the Compac Thread in the Hardware section:

We are looking at holding the contest this Friday and it will last 24 hours - Times to be announced.

I will post details as we come to an agreement about the rules and entry fee if there is one.

my 2c (repost from main):

yes, might be some modification, say, 8-12 hr continuous run and average speed reporting as a mark.
Somebody (AJRGale?) already reported some extreme readings for a few minutes.
I tried it before...compac gets tired of abuse after ~12 hr, but it could take it (to a degree) definitely longer than 10 min.

EDIT: OR, we can split competition into two parts: sprint (10 min) and 5000 meters (>8, <12hr)

[philipma1957]

400 freq for 24 hours is pretty easy with a big fan cooling 1 stick

I think I did 425 for 25 hours.

I have seen a few screens above freq 450.

[Mikestang]

I wouldn't be too impressed by any speed that only ran for 10 minutes.  I think in order to the speed to mean anything it has to be demonstrated that it can be maintained for more than a trivial amount of time - say 4 hours running at least.  I would think 24 hours would be an ideal test range, but if you're going to do a quick one then I would suggest several hours as a minimum to illustrate that the stick can hash at that rate/setup.

[RichBC]

There is a lot of options on what the contest could be, so perhaps over the coming Weeks we could have a few different ones? After all we need something fun to do with the Sticks.   A couple of thoughts.

I think it is good if the result could be measured centrally as opposed to relying on screenshots from the entrants. Not because of trust, if you want to cheat that is all to easy however things are measured, but because it makes for a level playing field and also enables everyone to see how people are doing. However do not know if this is practical so may have to rely on screenshots?

Am suspicious of any very short run measurement so suggest nothing less than 1 Hour, then perhaps 8Hr, 12Hr, 24Hr are good numbers?

Time differences may be a problem so might have to define a period during which runs must be done rather than an absolute start time?

Bottom line is lets do something and see how it goes.


Rich

[kilo17]

How about 2 fields like mentioned above.  I say a short 30 minute run and a 12 hour run.  I have a couple of ideas for cooling and unfortunately alcohol/DI doesn't last terribly long 

[kilo17]

I wouldn't be too impressed by any speed that only ran for 10 minutes.  I think in order to the speed to mean anything it has to be demonstrated that it can be maintained for more than a trivial amount of time - say 4 hours running at least.  I would think 24 hours would be an ideal test range, but if you're going to do a quick one then I would suggest several hours as a minimum to illustrate that the stick can hash at that rate/setup.

I kind of agree, usually for overclocking there is a short stress test performed - with CPU's it is things like SuperPi and a Prime95.  Both are fairly short runs.  I think hashing is as intensive as those programs and short runs of 10-30 minutes allow for extreme overclocking and say 12 hour runs for stability overclocking. 

[Jake36]

Guess I could always put good old #80 in (from the solo club pool overclocked contest) and see if it wants to do over 435MHz. Did have it start a couple of times at 465MHz before the contest, but couldn't get it to start that high reliably, so settled on 435.

Ran it for 60hrs, even though the contest was for 24hrs and with just the stock heat sink with fan for cooling.

But, I'll probably just watch this time.

[Guruclef]

im not entering but i really want to follow how people are going to cool there sticks in the competition

Hmm for a moment I fantasized about making use of the -80°C freezers in the university laboratory and getting creative. (Or the liquid nitrogen tanks??)

Then again, I'd better stay out of this one if I don't want to lose my stick and my job