AMD Zambezi news, info, fans !

[dess]

It does have a buffer so they are not stuck by the write speed.

It depends on the BW and size of this buffer...

And the l1 write is correct because you expect it to be in the facinity of the l2cache.

I think SW like AIDA64 could make a distinction of "short" (fitting in the WCC) and "long" (not fitting in the WCC) writes here. (I've noticed the authors about it.)

These results seem more like a read/write through up to the memory. So either a hardware or an Aida64 software issue.

Isn't it like to the L3, instead? Memory BW is roughly ~15 GB/s.

Anyway, the authors don't think it's AIDA64's fault. It could be a HW problem, but possibly they din't have had the latest AGESA when they've done tests on actual HW.

[duron]

AMD FX Processor Prices Lower Than Expected

On October 12, AMD will launch three new parts worldwide, the eight-core FX-8150, FX-8120, and six-core FX-6120, priced at US $245, $205, and $175, respectively.

http://www.techpowerup.com/152390/AM...-Expected.html

mmmm what does this mean?
PII prices will drop further?

[m_jones_]

I really hope they atleast come out before bf3 or im going to have to grab a cheap Phenom2 x4 to tie me over.

[-Boris-]

-Boris- what they tested was B2F and now we have B2G and who knows which Agesa code they had, because I saw the latest AGESA only with Asrock bios.

flyck nice correction

The pricing supports the hypothesis that BD sucks. You can't blame such low prices on bioses. AMD wouldn't sell them at that price if they outperformed 2600K. And an octa core with lower performance overall than an quad is nothing but a disappointment.

[Olivon]

The pricing supports the hypothesis that BD sucks. You can't blame such low prices on bioses. AMD wouldn't sell them at that price if they outperformed 2600K. And an octa core with lower performance overall than an quad is nothing but a disappointment.

Approx 300$ they said ...

[Pestilence]

what?

Approx 300$ they said ...

Blt and tank guys have the 8150 for 260 so 300 dollars was just an early estimate. I see it being around 240 on newegg

[-Boris-]

Approx 300$ they said ...

245$ for FX-8150 is the latest I've heard. 205$ for FX-8120 and 175$ for FX-6110.

[Pestilence]

Hmm I can speculate the 4100 being around 145 perhaps

[Lokinhow]

Hell, the FX-6110 is even cheaper than the X6 1100T.
While it's nice to have cheap chips, this doesn't sound good. Maybe FX6110 is slower than X6 1100T? I hope it's just AMD being (very) agressive on pricing

[Mats]

1 - DonanimHaber have been wrong before, and what's their source, anyway?

2 - As pointed out before, AMD have to set the prices lower than expected because Intel will lower theirs. See this as a precaution, AMD just don't want to lower the price the first thing they do after launch.

3 - You're comparing the 8150 with the price of the 2600K when in fact it will be the 2700K.

4 - AMD needs agressive pricing to attract customers? If AMD would have a CPU that's identical in terms of performance compared to an Intel CPU,
AMD's CPU would probably cost less. Edit: At least for now, it was a different story back in the days.

[muziqaz]

Besides, 8 core FX die is smaller than x6 thuban

[Ket]

Delys delays delays, thats all BD has ever been. I'm surprised anybody is even interested in BD anymore.

[-Boris-]

Besides, 8 core FX die is smaller than x6 thuban

Not much of a feat. And pretty irrelevant to the the end user, well except that large die sizes relative to power consumtion means better heat dissipation.

[muziqaz]

Delys delays delays, thats all BD has ever been. I'm surprised anybody is even interested in BD anymore.

I'm surprised you even bothered to come here and type this post in this thread.

[Mats]

Besides, 8 core FX die is smaller than x6 thuban

Hehe, well I'm curious what will happen to the X6 prices. Given that FX doesn't work in the majority of AM3 boards, maybe the X6 will stay pretty high, just like Athlon X2 939.
Long after AM2 was launched, the 939 CPU's were expensive.

BTW, have you seen how the board makers have backpedaled with the AM3 and FX support? Maybe old news, I dunno.

[wez]

4 - AMD needs agressive pricing to attract customers? If AMD would have a CPU that's identical in terms of performance compared to an Intel CPU,
AMD's CPU would probably cost less. Edit: At least for now, it was a different story back in the days.

I find to be the most believable scenario. Even if BD match or beats intels offerings, you'd think AMD need to get the word out. If that means low profit for the first few models, so be it. New steppings later on can bring price up, together with small performance boost to justify it.

Soon we'll know for sure

[Mats]

I'm surprised you even bothered to come here and type this post in this thread.

Ignore all drama queens.

[muziqaz]

Not much of a feat. And pretty irrelevant to the the end user, well except that large die sizes relative to power consumtion means better heat dissipation.

*sigh* If you reread from the post where specifically you were moaning about low price being the indicator of BD sucking, you would understand, that my post is not directed at end user. Beginning from your interesting(to say the least) comment up to my comment, the main idea of that mini discussion was why AMD might be pricing its chips at those rates?
One of the reasons might be its performance
Another reason is because of competition, which will lower its prices
And another reason is because they can as CPU die is smaller than x6 thuban. This does not concern end user, but this does concern AMD.
Even if x6 FX is cut down X8, it has lower die size than thuban, thus AMD will not be loosing more money than selling thubans.
We have precedent in this - AMD GPU department pricing their GPUs lower to get back the market share or to sell more GPUs, because they had smaller dies.
Though in this instance Intel 2600K has even smaller die, so I wonder what did AMD put into BD to have a larger die than 2600K(which has integrated GPU). But again we should count transistors of al things as there were some discussions around internets, that BD die has some empty spaces.

[Ket]

I'm surprised you even bothered to come here and type this post in this thread.

I'm making a very abundant point. Delay upon delay, no solid performance numbers, sketchy information. Surely you are capable of adding those things togther.

[muziqaz]

Hehe, well I'm curious what will happen to the X6 prices. Given that FX doesn't work in the majority of AM3 boards, maybe the X6 will stay pretty high, just like Athlon X2 939.
Long after AM2 was launched, the 939 CPU's were expensive.

BTW, have you seen how the board makers have backpedaled with the AM3 and FX support? Maybe old news, I dunno.

I think AMD will want to transition to BD dies as soon as possible. I guess we will hear thuban EOLs as soon as BD yields improve.
And low pricing, I do believe it is because they want to get back the market share.
P.S. what packpedaling? I have missed something I suppose.

[muziqaz]

I'm making a very abundant point. Delay upon delay, no solid performance numbers, sketchy information. Surely you are capable of adding those things togther.

Well, most of us here or anywhere else are capable of doing some add/mul by ourselves. We don't need yourself coming here and posting some some useless comment on how AMD is bad, and how they are screwed or oh no 2012 apocalypse is being pulled in into 2011 because amd delayed the release of their chip. We can see that they delayed BD, it means there some problems with it/ or something else. And I am sure they are doing everything they can do rectify those problems. I believe you are not perfect either, and most of the stuff you do does not come out the way you predicted couple of years ago. And please stop with this AMD-did-not-give-ME-info-on-how-BD-performs-thus-they-lost-ME-as-a-customer. Oh no!!! Don't let the door hit your arse on the way out :/

[Mats]

P.S. what packpedaling? I have missed something I suppose.

MSI and Asus had several old models that would work with Zambezi, but now they're not anymore.

I made this thread before, which shows 20 old models, but the lists from the manufacturers are changed, and only a few high end boards show no sign of Zambezi support.
MSI had 14 models, now they're 3.

[muziqaz]

MSI and Asus ahd several old models that would work with Zambezi, but now they're not anymore.

I made this thread before, which shows 20 old models, but the lists are gone from the manufacturers, and the individual boards show no sign of Zambezi support. I haven't checked all models tho.

haha, I remember saying that this move from mobo manufacturers was a bit strange and extremely interesting(from not good point of view). I even remember making a GIF with AM3 and AM3+ sockets flipping, thus showing that those sockets are not the same.
I myself have MSI 890FXA motherboard which was among one of those quoted to be compatible with BD. Yet flipping socket gif confirmed that it was impossible because you would need to cut out BD cpu pin out or cut the hole in the socket for that extra pin for BD.
I remember there was some rumours or even AMD announcement that they have some problems with 990FX chipsets, so at that time putting AM3+ socket onto 890 chipset based motherboard made sense. But now when it has been, what, 3 months since 990 mobos were released, there is no point in supporting two or 3 identical yet different lines of products.

[informal]

Dresdenboy has found some new Geekbench scores of 24C and 32C Opterons,submited 1 day ago!
Now compare this to the level of performance of Opteron ES 2x16C @ 2.1Ghz (1.4GHz shown on Geekbench score page is power management most likely).
Latest Opteron 2x16C @ 2.3Ghz result
Early Opteron ES 2x16C 2.1Ghz results

The difference in pure single and multithread performance is huge. Even if Turbo was a culprit for poorer single thread performance,still MT performance should be consistent . But we see massive difference,especially in MT floating point performance,ranging from 1.5x to 2.3x. Integer difference is also huge,ranging from 1.3x to 1.6x. Memory subsystem is also better on latest submission,but much less than what we see in pure integer and float performance. In my opinion ,the first ES opteron results had a silicon on the level of early zambezi B0/B1. BTW we see no difference with B2G/F when it comes to desktop but we see this massive difference when it comes to server parts which are launching shortly.

edit:
side by side comparison,in percentage
http://browse.geekbench.ca/geekbench.../484217/412187

[Mats]

Yet flipping socket gif confirmed that it was impossible because you would need to cut out BD cpu pin out or cut the hole in the socket for that extra pin for BD.

Looking at the socket alone won't tell you anything as long as you don't know what the CPU looks like. AM3 sockets looks different than AM2+ sockets, yet AM3 CPU's fits both.