Sneak peak at FX-8150 with SuperPi & wPrime

[BlueBlazer]

Sneak peek >> AMD FX series into mass production, and launch is expected in October [Exclusive].....

NordicHardware have now got hold of more concrete information on Bulldozers from one of its employees. We asked about what has actually been going in recent months with the FX series and why AMD has given us nothing but total silence. The answer was somewhat alarming, AMD has not even themselves know where they could launch the FX series, so they simply have not had anything to say to the media about it.

The situation remains very uncertain in AMD and you do not really know when the launch will take place, but mass production of AMD "Bulldozer" FX series has begun. As it stands today, hope AMD internally at a launch in 4-5 weeks. The latest rumors talk about a launch October 13 and in view of the 4 - 5 weeks we have heard directly from AMD, it may very well be true.

 

[ThatsABigOne]

20 seconds for Prime? Hmm..

 

[busydude]

Yup the same old stuff:

In the newer and more interesting wPrime 2:05 to get the CPU a final score of 15.815 seconds. We have reason to believe that this is not true, as Bulldozers in such cases would perform roughly equivalent to AMD A6-3650, according to our own tests . The odd, or should we say really bad performance figures may be due to several different things. But something that is likely is that the processor has either an older stepping, or the BIOS, the computer had not arrived with the latest AGES code from AMD.

Although performance figures are not the most credible, we can confirm one thing: If all goes as planned with AMD is launching the new FX series within 4-5 weeks. We should also then get the green light in on the new flagship from AMD has been worth the long wait. AMD still say they expect the FX series is the world's fastest consumer processor.

From the comments section:

Unfortunately, we had a short time with the system that we wrote did not have time to download more than what you see in the attached images in the article. Otherwise cpu-z has been high on the list to test blah. turbo and get more detailed info as what BIOS used.

As we wrote to old BIOS / AGES be the reason for the poor performance so it's just to take with a grain of salt. To my knowledge, only a ASRock motherboard with BIOS that has the right AGES or anyway the past?

About 4 - 5 weeks should the mystery be solved in all cases.

 

[BlueBlazer]

20 seconds for Prime? Hmm..

That's SuperPi result.

Yup the same old stuff:

From the comments section:

In 4-5 weeks, there could be a newer stepping. But then would another BIOS update be required for the new stepping? :hmm:

 

[drizek]

Asus released a new Sabertooth bios yesterday, but no change log. I haven't tried it yet.

 

[LOL_Wut_Axel]

I still wonder why they even bother to post SuperPi results. It uses the archaic X87 instruction set and is not good at all for disseminating single-threaded performance. It's bad for modern CPUs, even more for AMD's since it gives Intel's architecture a huge advantage.

Just to point out how bad SuperPi is, the old CPU in my laptop (Core 2 Duo P7450, now a T9600) overclocked to 2.7GHz calculates 1M in 19.465s, beating the FX-8150 by more than a whole second while being clocked 900MHz lower.

And yes, it was using SuperPi mod 1.5.

 

[T_Yamamoto]

pwned

i hate all the fake rumors -_-

 

[Arkadrel]

more fakes? jeez... ppl should just wait for the damn thing to get released and have real benchmarks from Anandtech and the like.

 

[sequoia464]

Asus released a new Sabertooth bios yesterday, but no change log. I haven't tried it yet.

Where did you find that bios? - I can only find the release 0402 bios.

 

[BlueBlazer]

more fakes? jeez... ppl should just wait for the damn thing to get released and have real benchmarks from Anandtech and the like.

This one comes directly from NordicHardware staff....

We also got the chance to briefly sit by a system equipped with 8-core processor, the FX-8150 with a clock speed of 3.6 GHz with a turbo mode at 4.2 GHz. Because of the very limited time period so we had time only to run the obsolete SuperPi 1.5 and wPrime 2:05. Unfortunately, the results far out to be something that represents the final product so take this information with a large pinch of salt.

 

[RussianSensation]

I still wonder why they even bother to post SuperPi results. It uses the archaic X87 instruction set and is not good at all for disseminating single-threaded performance. It's bad for modern CPUs, even more for AMD's since it gives Intel's architecture a huge advantage.

Actually, SuperPi performance scales almost perfectly with IPC between each of my Intel Core 2 Duo E6400, Q6600, Core i7 860 and i5 2500k!! I can't comment on AMD side, but SuperPi scaling shows very accurate linear IPC improvement across Intel's lineup in the last 5 years.

I wouldn't use SuperPi to necessarily infer real world performance in games, office apps, rendering, video encoding, etc. But if you strictly want to look at IPC improvements, it's one of the best benchmarks (assuming you don't skew the results with DDR3-2400 mhz memory, etc.).

If you look at Phenom I/II processors, they have 2 generations old IPC and SuperPi scores reflect that. Having said that, SuperPi may not be taking advantage of Bulldozer's specific architectural enhancements which will show up in real world apps and/or these benchmarks could be using borked BIOS.

This part of the news is more interesting: "If all goes as planned with AMD is launching the new FX series within 4-5 weeks."

So by the time BD launches, it'll be mid-to-late October! Wow, just 6-7 months away from IVB refresh. AMD better bring it to SB because even if they beat SB, Intel can turn up the heat quickly within 6 months with high clocked IVB chips, given how well SB scales. If FX8150 is the fastest consumer CPU in the market (and not just in encryption and rendering), I'll be uber impressed.

 

[LOL_Wut_Axel]

Actually, SuperPi performance scales almost perfectly with IPC between each of my Intel Core 2 Duo E6400, Q6600, Core i7 860 and i5 2500k!! I can't comment on AMD side, but SuperPi scaling shows very accurate linear IPC improvement across Intel's lineup in the last 5 years.

I wouldn't use SuperPi to necessarily infer real world performance in games, office apps, rendering, video encoding, etc. But if you strictly want to look at IPC improvements, it's one of the best benchmarks (assuming you don't skew the results with DDR3-2400 mhz memory, etc.).

If you look at Phenom I/II processors, they have 2 generations old IPC and SuperPi scores reflect that. Having said that, SuperPi may not be taking advantage of Bulldozer's specific architectural enhancements which will show up in real world apps and/or these benchmarks could be using borked BIOS.

This part of the news is more interesting: "If all goes as planned with AMD is launching the new FX series within 4-5 weeks."

So by the time BD launches, it'll be mid-to-late October! Wow, just 6-7 months away from IVB refresh. AMD better bring it to SB because even if they beat SB, Intel can turn up the heat quickly within 6 months with high clocked IVB chips, given how well SB scales. If FX8150 is the fastest consumer CPU in the market (and not just in encryption and rendering), I'll be uber impressed.

No, SuperPi is horrible for testing because it uses a very outdated instruction set that does not represent performance in any recent real-world program. X87 is old junk.

As far as launch and Bulldozer go, it's pretty obvious by now they're gonna ship for revenue in the very beginning of October at the latest. That means an official launch is possible by mid-to-late October.

I have a REEEALY hard time believing Bulldozer will be faster than Six-Core SB-E, which is okay. After all, it's priced a lot lower. At the same time, AMD wants their eight weak cores to compete against Intel's four strong cores. We'll see how it stands around a month from now, but I still think the same: significantly slower than Sandy Bridge single-threaded, but 5-10% faster in multi-threaded than the competing model (FX-8120 vs. i5-2500K). Given most applications use two-four cores, the i5 should be more balanced in typical desktop workloads.

 

[Despoiler]

Where did you find that bios? - I can only find the release 0402 bios.

On their FTP. version 705

ftp://ftp.asus.com.tw/pub/ASUS/mb/SocketAM3+/SABERTOOTH_990FX/SABERTOOTH-990FX-ASUS-0705.zip

 

[sequoia464]

On their FTP. version 705

ftp://ftp.asus.com.tw/pub/ASUS/mb/SocketAM3+/SABERTOOTH_990FX/SABERTOOTH-990FX-ASUS-0705.zip

Thanks

 

[blackened23]

Ivy Bridge won't increase performance more than 4-5% at most, per anands latest article. Most of the improvements are from higher clock speeds. IB isn't some amazing leap forward thats going to be magically 20% faster than SB.

http://www.anandtech.com/show/4830/intels-ivy-bridge-architecture-exposed/6

. Intel hasn't been too keen on pursuing clock speed for quite some time now. Clock for clock performance will go up by a small amount over Sandy Bridge

So if BD does turn out to be good and do somewhat well against sandy bridge, Ivy Bridge won't be a magical bullet that destroys it. It will have higher clockspeeds, lower power consumption, and better IGP. Thats about it. From what we've heard, the FX chips will be able to very high clockspeeds, so theoretically if BD matches SB, I don't think AMD will have problems raising clockspeeds to match IB.

Disclaimer, i'm neither an intel or AMD fanboy. I hope the FX chips turn out to be worthwhile, because that will be good for the entire communit....Anyway, i'm just dealing with theoreticals here. If BD does turn out to be good, IB isn't going to pose a big problem for it I don't believe.

Edit: jesus christ, those results are worrisome. Just read the entire page linked in the OP.

 

[996GT2]

These are not fake. It says right in the article that Nordic Hardware themselves tested the system. However, whether these results will be representative of the final product is up in the air...

 

[StrangerGuy]

My 5 year old 2.8GHz E6300 does 1M in <20 secs.

Unimpressed by AMD is an understatement.

 

[Will Robinson]

Unless the top SKU Bulldozer does 12 seconds or less for 1M Super pi 1.5 they are in deep trouble IMO.:\
My current CPU does it in around 10.8 sec.

 

[LOL_Wut_Axel]

My 5 year old 2.8GHz E6300 does 1M in <20 secs.

Unimpressed by AMD is an understatement.

Unless the top SKU Bulldozer does 12 seconds or less for 1M Super pi 1.5 they are in deep trouble IMO.:\
My current CPU does it in around 10.8 sec.

*sigh*

SuperPi is in no way, shape, or form an accurate comparison for Intel vs. AMD. Intel has decided even in their new CPUs to support this old and outdated instruction set, while AMD has not. That's why Intel's CPUs get much higher performance.

 

[996GT2]

*sigh*

SuperPi is in no way, shape, or form an accurate comparison for Intel vs. AMD. Intel has decided even in their new CPUs to support this old and outdated instruction set, while AMD has not. That's why Intel's CPUs get much higher performance.

I think you guys are forgetting that Nordic also tested wPrime...

...and this 8 core FX-8150 came out considerably slower than even a Thuban (which does wPrime32M in around 9 seconds).

 

[CHADBOGA]

My 5 year old 2.8GHz E6300 does 1M in <20 secs.

Unimpressed by AMD is an understatement.

I'm not impressed with AMD either, but it can't be because of how they go on Super Pi.

Super Pi is a useless benchmark for comparing AMD to Intel.

 

[LOL_Wut_Axel]

I think you guys are forgetting that Nordic also tested wPrime...

...and this 8 core FX-8150 came out considerably slower than even a Thuban (which does wPrime32M in around 9 seconds).

But then again, wPrime puts the Atom D510 as having the same performance as a Core 2 Duo SU7300 even though in the real world there's more than a 20% performance difference, even in multi-threaded.

 

[frozentundra123456]

Ivy Bridge won't increase performance more than 4-5% at most, per anands latest article. Most of the improvements are from higher clock speeds. IB isn't some amazing leap forward thats going to be magically 20% faster than SB.

http://www.anandtech.com/show/4830/intels-ivy-bridge-architecture-exposed/6



So if BD does turn out to be good and do somewhat well against sandy bridge, Ivy Bridge won't be a magical bullet that destroys it. It will have higher clockspeeds, lower power consumption, and better IGP. Thats about it. From what we've heard, the FX chips will be able to very high clockspeeds, so theoretically if BD matches SB, I don't think AMD will have problems raising clockspeeds to match IB.

Disclaimer, i'm neither an intel or AMD fanboy. I hope the FX chips turn out to be worthwhile, because that will be good for the entire communit....Anyway, i'm just dealing with theoreticals here. If BD does turn out to be good, IB isn't going to pose a big problem for it I don't believe.

Edit: jesus christ, those results are worrisome. Just read the entire page linked in the OP.

Hmmm, higher clockspeeds, lower power, improved IGP. Wow, no progress there.

 

[Ferzerp]

Hmmm, higher clockspeeds, lower power, improved IGP. Wow, no progress there.

Let's not forget the the transistor change brings about 20% faster switching at the same voltage. Yup, no way in the world that that plus the process shrink could result in substantially higher clock speeds. Nope, no way at all.

 

[inf64]

Ok we have a test system directly from AMD and it behaves like any other leaks. So facts are:
-zambezi 8C @ 3.6Ghz is slower than K10 in super pi <= not big of a deal so we can move on
-zambezi 8C @ 3.6Ghz is as fast as llano QC 3650 @ 2.6Ghz in wprime. This one is hard tp skip. This is a proof that none of the systems out there,even those that AMD show off to the press are representative of final performance. There is absolutely no chance that FX8150 is going to be slower than 2.6 QC K10.

The only logical conclusion is that all performance leaks shoot be disregarded until launch day.