E8400 + Folding = OMG...

[3.0charlie]

I've set aside temporarely my quest of finding the max. stable FSB of the Maximus Extreme / E8400 combo, and decided to start folding. And since I was too lazy to install VMware / Ubuntu the folding is done using dual Win SMP instances...

The E8400 is running at 9x445 (4.0GHz), while the Ram is at 760MHz (1520MHz).

(ppd: 4800... now, compared to my Q6600s - which are averaging 4760 ppd @ 3.6GHz...) ---> forget about those numbers, here are the real ones:

ppd: 2600 with dual SMP clients under Win, while the Q6600s are averaging 4760 ppd @ 3.6GHz w/ VMware, Ubuntu and dual Linux FAH clients.

(This is one amazing chip.) Well, it is but nowhere near what I thought I saw last night. I will surely install VMware this week-end and dual Linux instances to compare ppds.

Sorry for the huge mistake.

[enaberif]

Originally Posted by 3.0charlie

I've set aside temporarely my quest of finding the max. stable FSB of the Maximus Extreme / E8400 combo, and decided to start folding. And since I was too lazy to install VMware / Ubuntu the folding is done using dual Win SMP instances...

The E8400 is running at 9x445 (4.0GHz), while the Ram is at 760MHz (1520MHz).

ppd: 4800... now, compared to my Q6600s - which are averaging 4760 ppd @ 3.6GHz...

This is one amazing chip.

Dude! What does that say? The q6600s simply can't compete with the e8400 :D

[Eldonko]

How in the hell do you do that? I cant even do 2k ppd on mine.

[SugarJ]

How does he get that fast? 1 word. DDR3.

I was at 2950ppd under XP/Vmware/Xubuntu before I screwed up and moved to Vista 64.

[Eldonko]

Affinity prog + SMP x2? There is no way DDR3 more than doubles production.

[SugarJ]

I'm wondering why he's running 2 instances of the SMP kernel. I thought that the whole reason behind that program was to make use of the multiple cores so you don't have to run multiple instances.

Anyway, me poor old E8400 ticks over at 2460 ppd now at 9x445, same CPU clock speed as 3.0's.

[Eldonko]

Originally Posted by SugarJ

I'm wondering why he's running 2 instances of the SMP kernel. I thought that the whole reason behind that program was to make use of the multiple cores so you don't have to run multiple instances.

It is pretty much, but with the affinity prog you can run two instances of smp to trick it into thinking you have a quad. They finish slower but it is a good improvement over single.

[werty316]

Fah Smp Affinity Changer

Unless I am mistaken based on what I read in the description I thought you are supposed to run one SMP client for every 2 cores. I gave it try thinking this however I didn't see a boost in ppd. I think I'll try running two instances along with the SMP affinity changer to see if overall performance improves.

None the less that is one sweet ppd.

[stoanee]

Wow that is awesome! Wish I could afford to build a dedicated quad core rig coupled with that affinity program. It looks like your gaming performance will suffer though, so I will leave mine as it is.

[sswilson]

Originally Posted by stoanee

Wow that is awesome! Wish I could afford to build a dedicated quad core rig coupled with that affinity program. It looks like your gaming performance will suffer though, so I will leave mine as it is.

I just turn off FAH when it's time to game.

Great PPD 3.0charlie! The cache is doing it's thing, and you must have some wicked tight timings to go along with the high mem clockspeed..... What are your 1M SuperPi times like?

For folks looking at setting affinity for two VMware instances on a Quad.... I'm starting to believe the suggestions that XP doesn't always allocate the same "#" to each of the cores when it boots.... My PPD ranges from around 2500 PPD / instance to 2740 PPD / instance and it all seems boot dependant (If I'm getting 2700+ it will stay that way until I reboot, and then it's a $rapshoot between 2700 or 2500).