Bulldozer Performance Figures ( chases Sandy Bridge)

[seriousjack85]

AMD is Back!!! Beats SandyBridge.

http://cpuforever.com/Thread-Bulldozer-Performance-Figures-Outpace-Sandy-Bridge?pid=3970#pid3970

[Stupify]

This does look very good for AMD for now until Ivy comes out which when will cause another headache for AMD.
Nonetheless i rather wait for the actual product to be out and benched by various reviewers before I draw my conclusions.

[seriousjack85]

True. It has been long since AMD was on Top!

[DIREWOLF75]

Interesting to see the cachestructure.
Small L1 data for each core, and much larger shared ones for L1 instructions.
And 2M shared L2, that´s nice.

[Tiggerz]

Would be good if they could actually test non-engineering samples. After all, joe bloggs on the street is not going to have access to them are they.

[<R.t..>^>Fusionc]

Eh, in my view, any performance numbers coming in from ESs (engineering samples) are pretty useless since they don't represent the final retail product. Many of these early numbers (not specifically these) are also found to be faked as well. ESs are merely fore stability, compatibly, and validation testing; not for benchmarking.

[TAViX]

We need numbers from other independent sources....BTW, where are the tests??

[Stupify]

read the actual posts and they have the numbers in there.

[clone]

Eh, in my view, any performance numbers coming in from ESs (engineering samples) are pretty useless since they don't represent the final retail product.

so long as they are not faked they will only get better.

it's not like a car run on the dyno where the company later say's "oh well we changed the intake & exhaust for the sake of emissions hence the difference".

it's silicon, ES's are early versions that get improved upon later...... as mentioned at the beginning so long are they aren't faked they are indicative.

[Stupify]

After re-looking at those numbers, it seems that BD may not be the savior that AMD needed. It is a little here and there faster than 2600K and will only be 4-6Months earlier than Ivy Bridge, meaning AMD will just have a chip that can match but not thrash Intel even with dedicated hardware 8 threads unlike Intel's 8 HT threads. The bigger reason I am not impressed is because I am guessing the die size will be much bigger than Intel's SB and so Intel can easily squeeze AMD into selling at a lower ASPs and so further dimming AMD's future.

I hope AMD can truly merge the heterogeneous architecture of the CPU and GPU soon to be ahead - leave the FPU on the GPU and IPU (Integer) on the CPU and load up on cache and increase the bandwidth to CPU from RAM - more channels - rather dedicated channels to CPU and GPU.

[seriousjack85]

After re-looking at those numbers, it seems that BD may not be the savior that AMD needed. It is a little here and there faster than 2600K and will only be 4-6Months earlier than Ivy Bridge.

You must be Joking....That is big Achievement from AMD's point of View. At one point they were lagging behind INTEL now if they take the Crown for even 6 months is Big for AMD.

[TAViX]

read the actual posts and they have the numbers in there.

I see just some numbers, but I don't see any comparison tests or anything.....

[clone]

You must be Joking....That is big Achievement from AMD's point of View. At one point they were lagging behind INTEL now if they take the Crown for even 6 months is Big for AMD.

I agree with this, when AMD grows 100X larger and almost gets to be as big as Intel then I'll expect them to compete with Intel directly but the fact that their new cpu can beat todays parts is quite an achievement for them, given Intel's superiority on all things manufacturing it'll be no surprise that Intel will squeeze AMD again but at least things are getting better for consumers because of AMD's work.

Bulldozer to be honest is better than I expected....... if the numbers are true and real.

[Stupify]

I never said that BD was crap or it wasn't a good achievement for AMD. What I said was in terms of reality. If you look at the history - AMD had massive performance lead over P4 back in the early days of A64s yet they couldn't capitalize as much because Intel just squeezed them on the price perspective and their "rebates" (for exclusivity). Intel's behavior has left joe blow still thinking AMD is an inferior product regardless of the reality. Even with the help of OEMs and Microsoft, AMD did not deal such a massive blow to Intel back then with the x64-86. So even if AMD manages to match the performance would still give Intel the advantage over AMD, if not by a large margin. That's business and reality.

For what AMD has done and the direction they have taken with this whole Fusion approach, I applaud them and hope it does indeed give them the necessary advantage to at least capture the market share it has lost in past half decade or so.

[Cid]

Question is... Would you buy one?

[Stupify]

Happy with what i have already and don't think i will be buying one for another couple of years...

[Cid]

IF I was looking for a desktop I'd certainly consider it, I have better things to spend my money on for the moment as I already have a core i7 laptop which serves me well, I might buy a desktop later on when I have some money to spare and some time to have fun with it, not that I have a lot of either at the moment hehe
Forgot to mention, that price would have to be competitive but that is almost a given, maybe even from the launch date, we will see

[clone]

If you look at the history - AMD had massive performance lead over P4 back in the early days of A64s yet they couldn't capitalize as much because Intel just squeezed them on the price perspective and their "rebates" (for exclusivity). Intel's behavior has left joe blow still thinking AMD is an inferior product regardless of the reality

don't get me wrong I'm not dogging you but that period happened to be an exciting time in PC evolution that I happened to have been present for and I'm just enjoying the memory ....... by mentioning the period where AMD had an extended lead you have to consider all aspects that affected it.

1st off AMD did not have a "massive lead" over intel during the early P4 days, AMD suffered from poor mhz scaling and struggled to keep it's advantage, they were weak in motherboards and depended on 3rd party support which kept them out of the professional space, they also never gained a blatant lead until they went Dual Core, but during all of this time Intel was threatening OEM's and intimidated while leveraging their market position to limit AMD's attempt at making significant inroads worse still was the reality that AMD never had the capacity to get more market and were stuck making gradual inroads as they ramped as best they could.

the last is likely the primary reason why AMD only stole 9% of the market when they did dominate and on a side note also likely why AMD settled the legal dispute with Intel for a paltry 1 1/4 billion when it was clear Intel was guilty, AMD couldn't prove enough in damages because they were at maximum capacity the whole time.

that market share gain of 9% represented a 35% boost in AMD's production at the time and it was all they had.

[psycho_mccrazy]

I have just taken a look at these numbers... and the actual article is also linked in the xbitlabs frontpage story.

While the numbers do seem good compared to the Phenoms, it is now pretty evident that comparing core count to core count is also becoming a thing of the past, like clock speeds did before. Intel hyperthreading, AMDs modules, clock boost techs on both are all making the scenario more confusing. Was it the core count that made it happen? Was it the IPC? Was it the clock boost on lower threads?

While this may be OK for those comparing performance at stock speeds with the boosts enabled, it might be slightly misleading and confusing for those of us who overclock the entire thing and disable the boosts techs.

Also, one very big question remains unanswered about the Bulldozer. The Power Consumption! If this 4 module processor does outrank the i7 in stock performance in an overall manner, but is not able to do so with decent TDP, then all is not so rosy for AMD. And linked to the TDP also is the overclocking headroom!

[Hammer_Time]

Still waiting for retail Desktop Dozers and some real benchmarks...













They do have good marketing at least.