HyperThreading observations

[vaughan]

Here are the results of my HyperThreading comparison:

HT = ON

http://img93.exs.cx/img93/1820/sobhton2he.th.jpg (http://img93.exs.cx/my.php?loc=img93&image=sobhton2he.jpg)

HT = OFF

http://img93.exs.cx/img93/1107/sobhtoff0ld.th.jpg (http://img93.exs.cx/my.php?loc=img93&image=sobhtoff0ld.jpg)

Wow 11,000 cEMs/sec difference. I think its a con job.

[Empty_5oul]

lol, seems rather odd getting resaults like that.

all looks gd apart from the team being E3, whats with that ? also what processor was in there - its hitting some pretty high production rates.

[Beerknurd]

I think he has a P4 3.2....

With my P4 3.4 I can hit up to 3M with 1 instance, depending on the k/n pair.

[chaz]

Try 2 instances with HT Vaughan, mines faster that way.

[Beerknurd]

The speed isn't that much difference.... It is a bit higher though...

[chaz]

yeah its slight, but it helps.

Vaughan, I am crunching that very same test right now with my P4...odd.


oops, different n 8305517

[Empty_5oul]

lol, thay would be annoying if the same units were being processed again and again in a loop, lol

[Beerknurd]

As long as I got full credit I wouldn't care. Some tests are run more than once just to make sure a prime isn't missed.

[vaughan]

Empty_5oul team 83 = AMD Users.

I'll try two instances.

Yes its a P4 3.2E Prescott.

[Empty_5oul]

k, i thought there would be a reason behind it.

[Beerknurd]

there's always a reason...

[chaz]

lol, thay would be annoying if the same units were being processed again and again in a loop, lol

Yeah I make that mistake from time to time, reinstalling an os and just copying the DC project directory over from another pc. So I run the same wu over again, thats where those beauty - 300,000 numbers come from on sob. :oops:

[chaz]

BTW Beer, how the hell are you getting such good numbers with yours? I got my 3.0 at 3.36 and am still short. Maybe you got a Northwood, I read somewhere their shorter L2 pipeline makes up for the lack of cache. hmmmm

[Beerknurd]

I average about 2.8M to 3M. Depending on the k/n pair. I think right now it's running about 2.87M or so. My P3 733 is picking up the slack making me run about 3.15M. But other than SOB i'm not running any other projects except the key counting one's, and those don't use any resources. I have the SOB setting at normal and gave it a high priority in windows manager. And I have hyperthreading on. But other than that.... :?:

[chaz]

Ah I see, the extra 733mhz. That makes sense.

[Empty_5oul]

its nice to see the stats a bit higher on 17 again. Were doing pretty well and definatley slowing Free DC down though that are still out producing by 2x.

[chaz]

Not for long, i'm gonna give Vaughan a little push.... :twisted:

[vaughan]

:shock:

[Beerknurd]

:lol:

[Empty_5oul]

cant wait to see it chaz.
hopefully i should take ototero by the end of this week, then i only have to re-gain my position in RC5.

i notice dave has put some big 17 scores out in the past 24 hours http://marcc.no-ip.org/SoB/User.php?User_ID=7659

[Ototero]

I don't mind losing my position. I'm concentrating on other stuff :lol:

[AMDave]

it's a borged P4 2.8Ghz

[chaz]

:shock:

Dont worry, I dont think I can overtake you these days. Some of the faithful Bartons have gone bye bye.

[vaughan]

LOL, but I'm running CB, DHEP, SOB, some BOINC and I've started Grid again (to see if I can complete a task in it these days on an Athlon).

[Beerknurd]

why are you running DHEP?????? The stats have been down for weeks...... They haven't updated at all.

[Empty_5oul]

it does seem a waste when the stats pages are down as you dont know if all your processor power is going for anything.

[Beerknurd]

it does seem a waste when the stats pages are down as you dont know if all your processor power is going for anything.

my thoughts exactly....