Amd Officially Benches Bulldozer

[Manicdan]

most of you guys aren't reading into this right. the first intel chip is probably a 2500k because that is what amd is competing against. the second one is more expensive than a 2600k yet performs the same in that benchmark, so they chose the more expensive one to make themselves look better. but the real competition is a 2600k, and the data is representative of a 2600k, and that data is all we need to know.

this is not by accident. AMD CHOSE to clock and price these chips to compete well with the intel chips. they KNOW what their competition is, and WE KNOW what their competition is. you guys are spinning in circles, but the comparison is obvious. look clearly between the lines. either AMD competes well or they die, and they usually compete well. the implication of these benchmarks is that bulldozer competes well.

a 2500k is a great perf for its value, but the 2600k is not that much stronger in both single or multithreaded apps but much more expensive. so if AMD is just ahead of 2500k, and just behind 2600k, thats up to 100$ extra revenue per chip if they could get a little more perf out of it. practically every 1% increase in perf at this level could mean 2% more revenue

[RPGWiZaRD]

I just want to know whereabouts these maxes out for 24/7 use on air. Looks to be in the 5~5.5GHz or so range. One can always hope it's more matter of cooling/voltage you're pushing them with or the scaling physics nature of the architecture rather than other annoying limiting factors. The frequency WR gave us a pretty good hint though, now it just remains to be seen exactly how it scales with voltage.

[scottsche]

Ah yes, the famous ~5 minute random unknown movie Handbrake encoding benchmark. So much more useful than that ultra-obscure 3Dmark, Vantage or Super PI.

[BababooeyHTJ]

http://www.techspot.com/review/403-d...nce/page7.html


Do you still stand by that claim?

In games where I actually needed the cpu performance like Fallout New Vegas, Serious Sam HD, Borderlands, Darkplaces, and even Crysis my i5 2500k was a nice improvement over my i5 760. I didn't see a drop in performance in a single game that I tested.

[Lanek]

Something to do w/ the triple channel memory maybe...

nah it don't i had 15-20 fps less with CFX 5870 in Dirt2 with a 950 OC to 4ghz+ vs a 2600K (6vs4go)...., the memory have no impact in a game like that, specially at 2560x1600. Maybe the problem was coming from something else...

is really a so little difference in memory bandwith can make run a game slower or faster ?

[Spectrobozo]

nah it don't i had 15-20 fps less with CFX 5870 in Dirt2 with a 950 OC to 4ghz+ vs a 2600K...., the memory have no impact in a game like that, specially at 2560x1600. Maybe the problem was coming from something else...

indeed these results on Dirt 3 are really weird, the tests I found with Dirt 2 have the SBs with higher framerate, but this techspot review is the only one comparing CPUs on Dirt 3 I can find,

[Lanek]

indeed these results on Dirt 3 are really weird, the tests I found with Dirt 2 have the SBs with higher framerate, but this techspot review is the only one comparing CPUs on Dirt 3 I can find,

I have try find some other review too.. but i can't find one, i know what you mean. There's a nice amount of cpu scaling with Dirt2, but nothing for dirt3 and I7 1366 vs SB...

I have 2 950 and gigabyte's 1366 board here, but if i want test them with the same gpu, i will need completely unbuild my watercooling setup ( and install a win7 on a different HDD for a second PC ) a bit too much time and work just for check this.

[mockingbird]

AMD will have poor yields... Their low-nm fabs have not been proven yet...

[Chaserjzx100]

AMD will have poor yields... Their low-nm fabs have not been proven yet...

And you know this how?
also what are you talking about.. AMD don't have fabs they already sold it off.

[lm358]

TBH I really don't think Bulldozer is designed for "gamers" or really client business at all (that will be the future Fusion processors).

But I can't wait to get these chips in at work! They will slaughter the 2500/2600 (or the Xeon equivalents) under virtualization with 8-16 proper integer cores. (We don't run Cinebench or encode x264 all day, so floating point doesn't matter -- and is actually one area where HT can help a lot). Most real "business" work is all about integer power and that seems to be what AMD have targeted with the Bulldozer modules.

I guess that's why they are trying to use clock frequency (turbo)/unlocked multipliers to compete in the client space.

[SoF]

I hate these "cherry picked" benchmarks...why they don't have the balls to show some serious wprime or pifast or whatever really matters for benchers? also they compare a highly multi-threaded benchmark 8 vs. 4 cores....how stupid is that?
I smell something weird....this launch reminds me too much of Phenom II...there was a big hype, some good scores, then there came Intel with their new gen CPUs and all the hype went as quick away as the LHe...

€: I have a good marketing hint for Intel when I read the news again: Just claim all SuperPi WR's and show how bad AMD is there. Make some WR with a guy from Guiness (or 2 guys from Guiness - one beer guy, one record guy)..average joe will buy Intel for years then

[Greg83]

would it be ok to turn the hyperthreading off on the intel then for the tests to make it fair testing?

[Pestilence]

Microcenter already sells the 2500K at $180, so I want to see a 2600K at $200, that will be a bargain.

Has Intel stated what the 2700K is going to be? 6 Core with no HT? I can dream can't i?

Core I7 2700K - 6c/6t - 319.99
Core I5 2520K - 4c/8t - 219.99

Only in a Perfect World

HT vs no HT

[Hornet331]

I hate these "cherry picked" benchmarks...why they don't have the balls to show some serious wprime or pifast or whatever really matters for benchers? also they compare a highly multi-threaded benchmark 8 vs. 4 cores....how stupid is that?
I smell something weird....this launch reminds me too much of Phenom II...there was a big hype, some good scores, then there came Intel with their new gen CPUs and all the hype went as quick away as the LHe...

It was the phenom (I) launch where they pulled the 30k (or was it 60K duno anymore) 3dmark stunt...
It has some similarities, I give you that, but comparing a 8 core to 4 core when they cost the same is legit. Consumers don't care how many cores they get at a given price point. If AMD can deliver more core and thus more performance per $ thats all that counts. Still it seems absolute performance crown will still belong to intel.

[haylui]

I hate these "cherry picked" benchmarks...why they don't have the balls to show some serious wprime or pifast or whatever really matters for benchers? also they compare a highly multi-threaded benchmark 8 vs. 4 cores....how stupid is that?
I smell something weird....this launch reminds me too much of Phenom II...there was a big hype, some good scores, then there came Intel with their new gen CPUs and all the hype went as quick away as the LHe...

€: I have a good marketing hint for Intel when I read the news again: Just claim all SuperPi WR's and show how bad AMD is there. Make some WR with a guy from Guiness (or 2 guys from Guiness - one beer guy, one record guy)..average joe will buy Intel for years then

It doesn't work like this because even if people know Intel has all the WRs but the selling price is high. People couldn't afford even they know it is very good.

[informal]

I hate these "cherry picked" benchmarks...why they don't have the balls to show some serious wprime or pifast or whatever really matters for benchers?

Are you sure you want to see wprime from highly OCed Zambezi cpu? I think you might be surprised . Then I suggest a following :
€: I have a good marketing hint for AMD when I read the news again: Just claim all Wprime WR's and show how bad intel is there. Make some WR with a guy from Guiness (or 2 guys from Guiness - one beer guy, one record guy)..average joe will buy AMD for years then

Sounds familiar eh ?

[Mechromancer]

The only games I want to see a comparison of are ARMA II (OA) and Total War: Shogun II . Those games crush processors.

[JumpingJack]

I watched some "real" benches.. The AMD holds it's own very well against a 980X and when you kick in the price difference between it and a 980X it's a no brainer.
I'll say it this way, IF I was building a gaming machine I would use the FX8150 and not think twice about it, and buy the family's food for the month with the savings!

Were they all gaming benches?

[Hornet331]

The only games I want to see a comparison of are ARMA II (OA) and Total War: Shogun II . Those games crush processors.

*cough* Supreme commander*cough*

[ridney]

€: I have a good marketing hint for Intel when I read the news again: Just claim all SuperPi WR's and show how bad AMD is there. Make some WR with a guy from Guiness (or 2 guys from Guiness - one beer guy, one record guy)..average joe will buy Intel for years then

I don't think that's gonna work better than having a world record of highest clock frequency! That's what an average joe can always relate to

[Pestilence]

I don't think that's gonna work better than having a world record of highest clock frequency! That's what an average joe can always relate to

I really don't think Intel really cares tho. If they did they would drop a 6 core on 1155 and amd would be done with.

Do you hear me Intel? 6 core on 1155 and 8 core on 2011... We DEMAND it!!!!

[doompc]

@Pestilence, I don't see a 6 core on 1155 happening anytime soon.
As for 8 core on 2011, desktop X79 boards should work with Xeons...

[Pestilence]

@Pestilence, I don't see a 6 core on 1155 happening anytime soon.
As for 8 core on 2011, desktop X79 boards should work with Xeons...

Totally agree doom. Earliest 6 core on 1155 i see coming with Ivy Bridge.

[Glow9]

Benching a new chip against a chip that came out in 1Q of 2010.... Yeah that makes sense.

[qcmadness]

Benching a new chip against a chip that came out in 1Q of 2010.... Yeah that makes sense.

Still, 980x is the best multi-threaded CPU available in the market.