8 Packs 4ghz 32m Challenge win full Haswell 4770K Bundle.

[TrojanWhore]

What intra platform competition? There isnt one. We all know which cpu will win so it just comes done to a few people to tweak their memory and see who is best at tweaking (although that might still go to the person who spent the most money on memory but at least there is some skill there)

I meant the friendly 'bragging rights' rivalry springing up between some of the x58 + x79 users, who don't see themselves as having any real chance of winning the prize, but are competing between each other for fun.

on a seperate note, does anyone know how well threaded the workload is?
(say, for example, if I could gain access to a server at work, 16 cores 32 threads with 96GB RAM, or something there abouts! the cores would be far under 4GHz, but there are more of them!)

Edit: nvm, I was thinking of a different build, which supported multi cores, this is the single threaded benchmark isn't it!

 

[Greebo]

It loves more threads. Which is why a 16 core system should win it really even with slack timings. And I didnt see anything in the rules which said it had to be limited to 4/8 threads

 

[horrorwood]

Erm at no point does it say 32nm its a 32M comp ANY CPU can be used haswell or not

Ahhhh, saves swapping my CPU then sincy ivy is 22nm.

Now to order a haswell and have the potential of buy one get one free or not...

Hmmmm!

 

[Dyewitness]

Sorry don't see the point of this, a competition for people who already have high end kit to win more high end kit.

People that don't have a K processor are effectively ruled out.

Thats what i was thinking....

 

[Yaayuh!]

It loves more threads. Which is why a 16 core system should win it really even with slack timings. And I didnt see anything in the rules which said it had to be limited to 4/8 threads

It's uses a single thread, not multiple.

 

[Gambisk]

I ran this on my work laptop for a giggle 10:32 what do I win?

 

[Drangueos]

8 pack's screenshot is already faster than mine

 

[petridish]

I give up. Is it too complex to understand that it doesnt make sense to run a competition to win a high spec Haswell bundle for which the only possible winners will have high spec Ivybridge machines

Out of interest in what way doesn't it make sense? Do people who are fortunate enough to have Ivybridge machines not like competitions?

Anyway, does anybody know what a 'decent' time with 2500k and 8GB of 1600MHz RAM would be? I don't know if amount of RAM affects the time but I'd like to set myself a personal goal to beat

 

[wazza300]

this is the best I can do at 4ghz,no idea why must be the cpu clock speed

Spoiler

8m 3s

 

[subbytna]

I don't think my Q9550 is going to be able to compete even if I do crank it up to 4ghz. Might do it anyway when I get a spare momemnt to see what sort of time I get on my ageing chip XD

16 mins...dont' bother

 

[whyscotty]

Uploaded with ImageShack.us

Not the quickest

Will try again later


Damn - CPU speed wrong

 

[PiKe]

The only way to be competitive is to buy new gear... OH I GET IT, sneaky OCuK.

 

[wazza300]

The only way to be competitive is to buy new gear... OH I GET IT, sneaky OCuK.

just try it,its all down to your memory and cpu,so far ive seen a 2500k with bog standard ram beat my ivybridge cpu,so it's anyones guess

haswell will have the edge over current cpu's tho

 

[8 Pack]

Questions for 8 Pack

I would guess if you can't make the CPU speed of 4Ghz that being under would be OK ?

Are Haswell's disqualified ?
Seems odd that to win one you use one.

Haswells are fine. Yes under 4ghz is fine too.

I am glade so many are trying this. I will get a leaderboard up and running in the next couple of days.

 

[Akagi]

Out of interest in what way doesn't it make sense? Do people who are fortunate enough to have Ivybridge machines not like competitions?

That's not what I said.

This competition will be won by someone who already has a Haswell or Ivybridge based machine. Meaning that they will have little use for the prize.

Anyone who would truly appreciate the upgrade has zero chance of winning.

 

[Kaapstad]

just try it,its all down to your memory and cpu,so far ive seen a 2500k with bog standard ram beat my ivybridge cpu,so it's anyones guess

haswell will have the edge over current cpu's tho

Have do tried it without hyperthreading, it should not make much difference to the CPU efficiency but the temps may be a couple of degrees lower and gain you a few seconds.

 

[wazza300]

Have do tried it without hyperthreading, it should not make much difference to the CPU efficiency but the temps may be a couple of degrees lower and gain you a few seconds.

ive tried eveything

16gb/8gb of ram = same score
multithreading on/off = same score
various timings/voltages = roughly the same score

im losing 6seconds on the top score so far and that was with a 2500k and standard Kingston memory???

 

[Kei]

I think i've hit the limits of my chip. 1666MHz on the memory will not post even with very relaxed timings. NB doesn't want to exceed 2500MHz either and I don't want to go for suicide volts and roast the chip. Maybe it would be happier running just the kingston ram.

I'm seriously impressed that i've gotten this far considering it's a C2 stepping. Still no hope of winning anything, but it's been worthwhile seeing how far it could be pushed.

A small improvement nonetheless.

I want to see some more amd results, just so i have something to compare to.

 

[Greebo]

It's uses a single thread, not multiple.

Does it? Shows how out of touch I am then fromt he days I entered hwbot all the time. Always thought it ws multi threaded but you're right, it isn't. Must have been other benhcmarks that were then.

 

[Kaapstad]

ive tried eveything

16gb/8gb of ram = same score
multithreading on/off = same score
various timings/voltages = roughly the same score

im losing 6seconds on the top score so far and that was with a 2500k and standard Kingston memory???

I have just tried turning off 3 cores on my 3930k, turning off the hyperthreading and undervolting.

Did not make much difference.

What I don't understand is why at around 5.0ghz I can get the better of 2500/2600s but at 4.0ghz the best I can do is about 8.19 which is a lot slower. Does SB-E become less efficient at lower clock speeds or something.

I am not bothered about trying to win the competition (my entry had my ram running @1066mhz lol) but this 4.0ghz performance thing is a bit of a mystery.