AMD 'Bulldozer' gets an Update from Microsoft

[jhu]

Part of that problem is low availability, though. MSRP for the 8150 is $245, but they're scarce so retailers like NewEgg up the price.

Hmm, that still gives only 18 pps/$. I'll stick with what I have and see how Piledriver turns out.

 

[skipsneeky2]

Well there is another thread with someone claiming performance boosts in apps like cinebench with their 2600k due to the hotfix.

Perhaps any 2600k owners here could double verify the performance boost?

 

[wlee15]

Not that an hotfix documentation should be taken as official policy of Microsoft, but if Microsoft does treat an Bulldozer module as a core with SMT than that will substantially improve Bulldozer competitiveness in Microsoft SQL 2012.

 

[Magic Carpet]

So.. technically. What does this update really do and.. is it AMD only?

I don't have a windows system nearby to check the latter.

 

[frostedflakes]

Hmm, that still gives only 18 pps/$. I'll stick with what I have and see how Piledriver turns out.

I got 19.6 (assuming the same 4800 pps as the 2600K and that you could actually find an 8150 for $245). Of course you can't find them for $245 currently so the whole discussion is kind of moot. It's nobody's fault but AMD and GloFo's that they can't get their crap together and manufacture enough of these to meet demand.

No doubt the X6 (and the Phenom II in general) still offer great bang for the buck in some applications. If I did a lot of encrypting or video encoding w/AVX, though, BD would probably pull ahead of the X6 quite a bit in perf/$. Just depends on what you do. And as AMD can get the manufacturing issues worked out and more software is optimized for the new architecture, perf/$ will start to trend toward BD.

 

[sandorski]

Well there is another thread with someone claiming performance boosts in apps like cinebench with their 2600k due to the hotfix.

Perhaps any 2600k owners here could double verify the performance boost?

1 performance boost and 1 Blue Screen. Although he/she claimed the overclock was the cause of the BS.

 

[Chiropteran]

Sounds more like yield issues and they're turning off 2 modules.

Can you pay attention to context please? I was responding to skipsneeky2 who said "over abundent fx cpus stores can't even give away", which is obviously a lie if the biggest online retailers are sold out.

Yield issue, irrelevant, point is the CPUs are not "over abundent" (sic) if they are sold out.

 

[jhu]

I got 19.6 (assuming the same 4800 pps as the 2600K and that you could actually find an 8150 for $245). Of course you can't find them for $245 currently so the whole discussion is kind of moot. It's nobody's fault but AMD and GloFo's that they can't get their crap together and manufacture enough of these to meet demand.

No doubt the X6 (and the Phenom II in general) still offer great bang for the buck in some applications. If I did a lot of encrypting or video encoding w/AVX, though, BD would probably pull ahead of the X6 quite a bit in perf/$. Just depends on what you do.

True enough. Unfortunately, povray doesn't have explicit support for AVX. It doesn't even have real support for SSE2 either other than being able to use more XMM registers to calculate 64-bit floats (16 XMM vs. only 8 ST(*) registers). Hmm... I think I'll look into possibly thinking about changing that.

 

[jhu]

Can you pay attention to context please? I was responding to skipsneeky2 who said "over abundent fx cpus stores can't even give away", which is obviously a lie if the biggest online retailers are sold out.

Yield issue, irrelevant, point is the CPUs are not "over abundent" (sic) if they are sold out.

He never mentioned which fx CPUs. He just said overabundant fx-4100 can't be given away at Fry's. You mention the 8150 being sold out, which it has been on Newegg for quite a long time. That points to yield issues (which have been confirmed on other sites), which likely points to harvesting those into lower spec parts (eg 4100). fx-4100 and fx-6100 are abundant and easily obtained.

 

[Chiropteran]

Pentium Pro was intended for the workstation market. AMD's FX line is intended for the rest of us, who actually run Windows 7 (well not me anyway). If Bulldozer were Opteron-only and then AMD came out with an optimized version of Bulldozer for the consumer market, then your objection would be valid. However, AMD didn't do that. It's the same architecture for Opteron and FX.

Why are you changing the question? I was responding to THIS:

So AMD makes a processor that doesnt perform at 100% on the most popular OS out there? And that seems ok to you?

So Intel makes a processor that doesnt (sic) perform at 100% on the most popular OS out there? And that seems ok to you?


Yes, it seems okay.

Everything else you bring up is irrelevant to the question. Also, factually false.

"The Pentium Pro is a sixth-generation x86 microprocessor developed and manufactured by Intel introduced in November 1, 1995 [1]. It introduced the P6 microarchitecture (sometime referred as i686) and was originally intended to replace the original Pentium in a full range of applications"

But that chip was expensive and intended for workstations in the beginning.

How does being expensive make it any better? You won't buy a $200 CPU because it's not perfectly optimized for the most popular OS, but you will fork over $1500 for a similar chip which is equally unoptimized? What kind of logic is that?


As far as the intentions, you are wrong. Until it flopped, Intel hoped that the Pentium Pro would simply replace the Pentium in all applications across the board. See the quote above.

 

[skipsneeky2]

He never mentioned which fx CPUs. He just said overabundant fx-4100 can't be given away at Fry's. You mention the 8150 being sold out, which it has been on Newegg for quite a long time. That points to yield issues (which have been confirmed on other sites), which likely points to harvesting those into lower spec parts (eg 4100). fx-4100 and fx-6100 are abundant and easily obtained.

The fx 4100 is their budget chip but the x4 955 is still a better buy and better overall chip.

The other chips don't make sense for fresh builds perhaps major upgrades if phenom 2 options are sold out or are overpriced.

I almost considered one myself but the reviews were pretty awful i always base my purchase decisions off good reviews and how many reviews were bad?

Plenty

 

[jhu]

How does being expensive make it any better? You won't buy a $200 CPU because it's not perfectly optimized for the most popular OS, but you will fork over $1500 for a similar chip which is equally unoptimized? What kind of logic is that?

Intentions and work loads are important. That $1500 Interlagos is going into servers where multithreading is so common that it makes little if any difference with regard to the thread scheduling for the processor. Plus they're multiprocessor capable (try finding a 4-socket motherboard for Zambezi; you can't because it doesn't exist). For an n module chip, it only matters if there are less than 2n-1 threads that the processor is working on (which is more common for consumer use).

As far as the intentions, you are wrong. Until it flopped, Intel hoped that the Pentium Pro would simply replace the Pentium in all applications across the board. See the quote above.

In some sense it did when they made the Pentium 2. But really, Intel had the same intentions for Itanium. That doesn't change the fact that both Pentium Pro and Itanium were way more expensive than their consumer level processors and didn't run consumer level software optimally. The introduction of Pentium Pro was the point where Intel began separating its marketing.

 

[Vesku]

Would be great if Bulldozer was the Pentium Pro of their future CPU line.

 

[jhu]

Would be great if Bulldozer was the Pentium Pro of their future CPU line.

If "heavily multi-threaded" is the new "32-bit computing" then AMD should be OK as long as they can remain a financially viable entity.

 

[bononos]

...... With BD, you've got a slower processor and OS performance issues. Once the OS issues are fixed, typical performance will still only be about as good as it is in Linux, where the single-threaded performance is still about the same as in Windows. If the typical single-threaded performance, compared to a Phenom II, had improved significantly, the patch would still be needed, and still would be made, but you wouldn't care. It would be gravy to enhance good mashed potatoes, instead of gravy to flavor KFC "mashed potatoes".
......

Problem is that the colonel's boys have eaten so much KFC that their tastebuds have atrophied and KFC mash feels superior to anything else out there.

 

[Stoneburner]

Looks like the patch makes BD show up as 4C/8T in Windows now instead of 8C/8T. The marketing guys at AMD just cringed.

you mean marketing guy (singular). They fired everybody except the dude that replaces notepads and staples.

 

[Olikan]

BF3 is a little different. You can see the top ss is before the patch. The bottom one is from after patch. There is a change in core utilization. I got a 15-20fps increase on battlefield 3 with my CF and eyefinity.went from 30-40 to 50-60. MW3 increased by 35-45fps for me too. Below shows MW3. First is before, second is after patch.

OMG!
i just can't belive this

http://www.overclock.net/t/1185039/the-1st-windows-7-scheduler-patch-results-are-in-have-fun/60#post_15955453

 

[Edrick]

So Intel makes a processor that doesnt (sic) perform at 100% on the most popular OS out there? And that seems ok to you? Yes, it seems okay. Everything else you bring up is irrelevant to the question. Also, factually false. "The Pentium Pro is a sixth-generation x86 microprocessor developed and manufactured by Intel introduced in November 1, 1995 [1]. It introduced the P6 microarchitecture (sometime referred as i686) and was originally intended to replace the original Pentium in a full range of applications"

As IntelUser already stated, the PPro was not designed for the desktop market. Which makes your counter point invalid. Please use a different example.

Also, for AMD to release a CPU that works best on a OS that is a year away (Win8), makes no sence to me. It seems to be a very poor business decision. And as an earlier poster stated, AMD clearly had copies of Win7 to test with, so what happened?

The point is that Win7 (and XP) make up the majority of users the BD was targeting, and the CPU does not perform as well as it can on these OSes.

 

[Edrick]

How does being expensive make it any better? You won't buy a $200 CPU because it's not perfectly optimized for the most popular OS, but you will fork over $1500 for a similar chip which is equally unoptimized? What kind of logic is that? As far as the intentions, you are wrong. Until it flopped, Intel hoped that the Pentium Pro would simply replace the Pentium in all applications across the board. See the quote above.

How many people did you know that used PPro's in their desktop? And the few who did were not using Win95 I will tell you that.

And I assume you worked for Intel back in 1995 as you seem to think you know about their marketing stratergy so well. I never recall the PPro being marketing at desktop users (the Pentium MMX was).

 

[sandorski]

OMG!
i just can't belive this

http://www.overclock.net/t/1185039/...atch-results-are-in-have-fun/60#post_15955453

Interesting. If true, that's significant.

 

[Idontcare]

Also, for AMD to release a CPU that works best on a OS that is a year away (Win8), makes no sence to me. It seems to be a very poor business decision. And as an earlier poster stated, AMD clearly had copies of Win7 to test with, so what happened?

I suspect AMD knew well in advance that the scheduler needed to be patched in order to maximize performance.

But what good would it have served them to have gone public, bitching and moaning that Microsoft wasn't internally prioritizing and escalating the scheduler patch on a timeline that was going to intersect Zambezi's launch?

Who wants to declare to the world that they don't command enough priority at Microsoft to get critical patches done in a timely manner?

And look at it from Microsoft's perspective. They use to go to efforts to be compatible with MIPS, DEC Alpha, and Itanium...but one by one they dropped those efforts as the respective return on investment fell along with marketshare.

What incentive is there for Microsoft to worry about patching their scheduler when the marketshare of Zambezi owners is probably 1% or less?

They were going to get around to it when they got around to it, and AMD had to just sit idly by and wait for their gift horse without making it too obvious to the world that they were the beggars in the equation.

 

[frostedflakes]

I also wonder if maybe they didn't see the patch as a huge priority early on because they were expecting BD clocks to scale much higher and for turbo to compensate for sub-optimal scheduling. Then whenever the scheduler patch was released, it would just be gravy, a free 10-15% performance improvement on top of already decent single and lightly threaded performance. Remember that Anand said AMD was hoping for 30% higher clocks than their previous architecture. If you use Phenom II X6 1100T (3.3 base 3.7 turbo) as the baseline, that would put BD at about 4.3GHz base and 4.8GHz max turbo. But then when BD wasn't released with turbo anywhere near 4.8, more optimal threading was necessary just to eke out halfway respectable single and lightly threaded performance from Zambezi.

 

[sandorski]

I suspect AMD knew well in advance that the scheduler needed to be patched in order to maximize performance.

But what good would it have served them to have gone public, bitching and moaning that Microsoft wasn't internally prioritizing and escalating the scheduler patch on a timeline that was going to intersect Zambezi's launch?

Who wants to declare to the world that they don't command enough priority at Microsoft to get critical patches done in a timely manner?

And look at it from Microsoft's perspective. They use to go to efforts to be compatible with MIPS, DEC Alpha, and Itanium...but one by one they dropped those efforts as the respective return on investment fell along with marketshare.

What incentive is there for Microsoft to worry about patching their scheduler when the marketshare of Zambezi owners is probably 1% or less?

They were going to get around to it when they got around to it, and AMD had to just sit idly by and wait for their gift horse without making it too obvious to the world that they were the beggars in the equation.

I find the whole discussion of whether this Patch means AMD screwed up their Design or not to be silly. Even the X2 required a Patch to work properly. Just seems to be nitpicking IMO.

 

[IntelUser2000]

There's a HardOCP review of SLI and XFire comparions between 8150 and the 2600K. The FX-8150 is performing pretty bad.

http://www.hardocp.com/article/2011/11/03/amd_fx8150_multigpu_gameplay_performance_review/2

Looks like it might be worse than Phenom II X6. The patch might make it so it at least makes sense as a CPU.

 

[soccerballtux]

so, how in the world didn't AMD get Microsoft to implement this by launch date? Failure...