September 22nd ETA for AMD FX processors

[frozentundra123456]

To a point... see my above stream of consciousness.

My feeling is that single thread performance at the moment is more than good enough in mainstream CPU's that the biggest gains for the consumer are in the entire industry moving towards multiple threads and scalability as the standard.

I wonder if thats AMD's long term game plan- a shift towards the bottom end and catering to users that don't need a ferrari, but could sure do with a pickup that will get them from A to B fast enough and still tow a trailer behind it.
Build it and they will come- software is only going to take more advantage of more threads as time goes on.




I miss the sound of my 56k wub

Ummm..... hate tell you, AMD cant shift toward the bottom end, they already are the bottom end.

 

[StrangerGuy]

Ummm..... hate tell you, AMD cant shift toward the bottom end, they already are the bottom end.

Low-end? AMD now doesn't even have a low-end to speak of. Pentium G620 is superior than the Athlon II X2 in every way AND cheaper to boot. It's scary how Intel is pwning AMD from bottom to top.

 

[Arkaign]

Directed at nobody in particular, but noteworthy :

Yes, more cores are always good. But it's not a magic bullet. You still need the ability to wail out very high performance PER core for overall great performance when a particular app gets thirsty for power.

Now the biggie :

If you run windows 7, bring up task manager, now hit the performance tab, then open 'resource monitor. LOOK at all of the processes! Most of them don't do all that much a lot of the time, but damn if there's not a ton of them anyway. If you have a single-core system, does that mean you can't run windows? Of course not, though single-cores do suck in today's world. BUT, that core can still juggle many threads just due to windows process scheduling. This is why you want the best balance possible between lots of cores (to keep multitasking smooth with lots of crap running), AND good single-core IPC/performance, so when you fire up something that needs a ton of performance in only 1-2 threads, that you've got the firepower to fight it.

Think of it as a fire department fighting a large fire with smaller fires here and there. 8 tiny garden hoses will be less effective than 4 big badass 100 gallon a minute nozzles.

 

[sangyup81]

Why all the hang up on single threaded performance anyways? 2 threaded, 3 threaded, and 4 threaded applications make up the majority of what is used and even many the 3 and 4 threaded apps are very inefficient with the 3rd and 4th threads.

 

[Arkaign]

Why all the hang up on single threaded performance anyways? 2 threaded, 3 threaded, and 4 threaded applications make up the majority of what is used and even many the 3 and 4 threaded apps are very inefficient with the 3rd and 4th threads.

The reason is why you see i3 2100 often faster than Phenom II X4 in games that DO use 4 cores. Efficient load-balancing is next to impossible. Most games that use more than 2 cores still use mostly one core, with dramatically less load on the other cores. Hence, you want that one mostly loaded core to really fly. On apps that efficiently load all the cores (encoding), it's much less of a problem.

Nobody sane is saying (at least I don't think), that it's important to have a fast single-core processor. What we're saying is that single-threaded performance is important because many slow cores are worse than less fast cores for many situations.

Even so, I don't recommend dual-cores any longer for most users. Quads are cheap and are far more rounded for all-around multitasking performance. If we can get more cores while still increasing single-threaded performance, all the better.

What we don't want to see is a move towards 6 or 8 cores, but none of the cores are capable of being competitive with the single-threaded performance of CPUs like SB. At least not in the gaming/enthusiast sector. For servers and general desktop derping it's fine.

 

[sangyup81]

The reason is why you see i3 2100 often faster than Phenom II X4 in games that DO use 4 cores. Efficient load-balancing is next to impossible. Most games that use more than 2 cores still use mostly one core, with dramatically less load on the other cores. Hence, you want that one mostly loaded core to really fly. On apps that efficiently load all the cores (encoding), it's much less of a problem.

Nobody sane is saying (at least I don't think), that it's important to have a fast single-core processor. What we're saying is that single-threaded performance is important because many slow cores are worse than less fast cores for many situations.

Even so, I don't recommend dual-cores any longer for most users. Quads are cheap and are far more rounded for all-around multitasking performance. If we can get more cores while still increasing single-threaded performance, all the better.

What we don't want to see is a move towards 6 or 8 cores, but none of the cores are capable of being competitive with the single-threaded performance of CPUs like SB. At least not in the gaming/enthusiast sector. For servers and general desktop derping it's fine.

Exactly what I'm trying to get at! The only application I've ever seen benefit from 4 cores vs. 3 is Adobe Flash based on results from Flash Benchmark 08. You won't recommend Dual Cores but how about 2-core, 4-threaded CPUs like the i3, especially in laptop applications?

Personally, I may actually go with a BD since I have a compatible MB already and since I crunch on BOINC. Of course this is going to be dependent on the performance of those new flexible 256-bit FPUs

 

[Arkaign]

Yeah, to clarify, definitely dual-cores with or without HT make sense for plenty of notebook users, as 90%+ of notebooks are garbage for gaming anyway. Better battery life and decent performance for common tasks is fine.

 

[Ancalagon44]

You wrote nothing to respond too. The fact you think something which may be poor as single thread will suck as multi shows you don't really understand multi threaded at all. How well something scales makes a huge difference when you start adding multiple cpus and even when concerning gpus as well.

This is not about AMD or Intel for that matter. You can babble all day, but the fact remains we are in the computing world a multi-threaded future. I'll take a cpu that has 80-85% single threaded performance if I can have two of them working together over the faster single core. And all 486 type posts are just plain silly.

Everyone in this entire thread is telling you that you are wrong and understand nothing about multi threading but you still carry on? Just admit it, you dont know anything!

Something poor at single threading is going to suck at multithreading! Of course it is. If not, then my example of making a multi core 486 would be the best idea ever. We could have 128 core 486's. Why would you not want to do that? Come on, tell me. You seem to think that somehow multithreading makes individual core performance irrelevant - it does not. What you are saying is that individual cores are irrelevent, thus we need as many cheap cores as possible. Not how the world works.

 

[Arzachel]

Think about it this way: if you play a game/use an app that uses 4 threads or less, a FX 8150 instamagicaly turns into a ~i5 760@3.5ghz(Depending on Bulldozers IPC) due to turbo.

What AMD base their strategy on, in my opinion, is that applications, that utilize 4 cores or less, won't be as CPU hungry thus you won't be bottlenecked. If an application uses more than 4 cores, then AMD can show its multi-threaded superiority.

Example: would a i5 760@3.5ghz bottleneck you much more than i5 2500k at something like BF:BC2? Probably not. Would FX 8150 be better than i5 2500k at BF3, which can use up to 16 threads? I'd say yes.

 

[BlueBlazer]

Think about it this way: if you play a game/use an app that uses 4 threads or less, a FX 8150 instamagicaly turns into a ~i5 760@3.5ghz(Depending on Bulldozers IPC) due to turbo.

What AMD base their strategy on, in my opinion, is that applications, that utilize 4 cores or less, won't be as CPU hungry thus you won't be bottlenecked. If an application uses more than 4 cores, then AMD can show its multi-threaded superiority.

Example: would a i5 760@3.5ghz bottleneck you much more than i5 2500k at something like BF:BC2? Probably not. Would FX 8150 be better than i5 2500k at BF3, which can use up to 16 threads? I'd say yes.

Games don't always behave like that even when multi-threaded (something that I and Arkaign has already pointed earlier, on how not all cores are evenly used). Another thing is data dependancies (or thread dependancies) where one thread had to wait for another thread to complete, or one thread waiting for data passed from another concurrent thread. This is also common in games.

 

[classy]

Everyone in this entire thread is telling you that you are wrong and understand nothing about multi threading but you still carry on? Just admit it, you dont know anything!

Something poor at single threading is going to suck at multithreading! Of course it is. If not, then my example of making a multi core 486 would be the best idea ever. We could have 128 core 486's. Why would you not want to do that? Come on, tell me. You seem to think that somehow multithreading makes individual core performance irrelevant - it does not. What you are saying is that individual cores are irrelevent, thus we need as many cheap cores as possible. Not how the world works.

First of all unlike most I support, daily over a dozen developers, a finance department that runs large number crunching excel sheets, and bunch of other people who use the internet with 5 different messengers and a lot ftp work, with virtualization as well, locally. I daily see people who truly use the internet for business. You and most others speak from your perspective of your house. I did not say single core performance is irrelavant, what I said was the concept is a dead horse and it is. If it wasn't you would still see single cores being pushed out, but you don't, not even in cell phones.

We are at a point where scaling to be able to run more programs at the same time is more productive than being able to run 2-3 faster. I can tell you flatout an older quad core 6600 is better than a higher clocked newer dual core. Fact not fiction, the company I donate pcs to is very happy though . I am not saying build a slow core, but what I am saying is moving in the direction of scaling, even at the expense of a slightly slower single core in my opinion is better.

Single core is not a new argument, I been around a long time. It was the same argument used back before multi core processors even existed or even thought would be possible, especially on this scale. As for gaming, I have said for years having a game run smoother was a better experience than having it run faster, but with a bunch of stutter.

Now to the rubber meets the road. Truth be told, the comment is being made because the Intel boys are now worried that AMD does have decent product on its hands. So as way to deflect and maybe avoid having to eat some crow, now its well its still slow in single thread. Thats all this is about. I am an Intel guy first, but I have AMDs running in my home as well, but truth be told the argument is just about let me see what Intel vs AMD still is better at because it looks like AMD is going to bring some heat . And at the end of the day competition is better for us all. When you have several hundred thousand dollars you spend yearly, some of us like to get more for less.

 

[ransomlist]

This could be oopening a can of flesh eating killer worms here, but lets consider this-

If AMD was selling BD as a quad core with 'an alternative to weak hyperthreading' and the 3rd party marketing bumpf was discussing it as such, then how would this effect the discussion here?

 

[ransomlist]

Games don't always behave like that even when multi-threaded (something that I and Arkaign has already pointed earlier, on how not all cores are evenly used). Another thing is data dependancies (or thread dependancies) where one thread had to wait for another thread to complete, or one thread waiting for data passed from another concurrent thread. This is also common in games.

A lot of us are waiting to see if DICE has pulled it off at the moment, the Alpha for BF3 scaled very well indeed from dual-quad-hex-octo (HT) and they have made a point of stating thats its optimized for multithread and HT performance.

Regards BD, i dont have the background to try and make an informed guess as to what shared resources means for this- but i'm okay with stating that as a consumer, i probably wont actually notice a difference between BF3 with an i5-2500K/i7-2600K/FX8150 because with a good card they will all perform more than adequately and wont be the bottleneck on visible performance on my 60hz screen.

 

[sequoia464]

...care to speculate??

 

[Elixer]

...care to speculate??

No need to speculate.
They are not allowed to show dates as per NDA.
It obviously was put up in error, since they don't want to be the last one getting AMD's new chips (if they would even get them at all).
This is the same reason the egg don't list the chips, but, obviously, they know when they will get a shipment in, if they haven't gotten a shipment already.

 

[sequoia464]

No need to speculate.
They are not allowed to show dates as per NDA.
It obviously was put up in error, since they don't want to be the last one getting AMD's new chips (if they would even get them at all).
This is the same reason the egg don't list the chips, but, obviously, they know when they will get a shipment in, if they haven't gotten a shipment already.

Kind of figured that it was posted in error - surprised that the listings stayed up as long as they did, I'm just wondering about the chance of the chips actually being available for sale on the 22nd.

 

[BlueBlazer]

...care to speculate??

Just noticed, they pulled out the listings.

Kind of figured that it was posted in error - surprised that the listings stayed up as long as they did, I'm just wondering about the chance of the chips actually being available for sale on the 22nd.

Hopefully that is still slated for 22nd. Unless there is some other unforeseen issues with shipping, non-delivery, change of product notification, product embargo or more delays?!

 

[nenforcer]

Just noticed, they pulled out the listings.

Hopefully that is still slated for 22nd. Unless there is some other unforeseen issues with shipping, non-delivery, change of product notification, product embargo or more delays?!

I find September 22nd to be completely unrealistic. We would have heard a lot more advertising wise from AMD and not just overclocking benchmarks if that were to be true.

NordicHardware claims is October 13th now and they have some photos from a German computer taken with some benchmark results (SuperPI, wPrime).

AMD won't just dump it on the market without letting everyone besides the enthusiasts knowing about it.

 

[BlueBlazer]

I find September 22nd to be completely unrealistic. We would have heard a lot more advertising wise from AMD and not just overclocking benchmarks if that were to be true.

NordicHardware claims is October 13th now and they have some photos from a German computer taken with some benchmark results (SuperPI, wPrime).

That should be in-line with the mid-October speculations and predictions earlier (including the AMD FX giveaway contest clues).

AMD won't just dump it on the market without letting everyone besides the enthusiasts knowing about it.

Its worrisome that AMD has been keeping very quiet lately, even no announcements on shipping dates/schedules this close to speculated/predicted "launch" deadline (less than a month away from mid-October). I hope that does not mean more delays. :hmm:

 

[Elixer]

Its worrisome that AMD has been keeping very quiet lately, even no announcements on shipping dates/schedules this close to speculated/predicted "launch" deadline (less than a month away from mid-October). I hope that does not mean more delays. :hmm:

It is true that you, me, and most of the other forums around really hate AMD not letting more details out, but, it wouldn't be unheard of for them to have 0 heads up about the release date. Interlagos was announced on September 7, and nobody had a heads up as far as I can tell. And that date was a Wednesday.

Since the folks @overclock.net will have a BD streaming event at or pretty close to the launch (and they said it will be on A Saturday) it is sounding like a Monday release.

It is all in marketing hands now...

 

[Arg Clin]

ETA 22/9 and no beahcmarks yet. I'm feeling increasinly confident that AMD has taken the approach to catch Intel by surprise, and keep a tight lip untill BD is availible in the marketplace.

 

[Tanclearas]

ETA 22/9 and no beahcmarks yet. I'm feeling increasinly confident that AMD has taken the approach to catch Intel by surprise, and keep a tight lip untill BD is availible in the marketplace.

<opens fortune cookie>
You are going to be very disappointed.

 

[JimKiler]

ETA 22/9 and no beahcmarks yet. I'm feeling increasinly confident that AMD has taken the approach to catch Intel by surprise, and keep a tight lip untill BD is availible in the marketplace.

Do you remember all the previews we had for Brazos before it was launched and almost nothing and zero benchmarks from AMD prior to BD launch. We have to hope AMD can correct whatever issues BD has for the second gen and onward otherwise all us enthusiasts will be intel only.

 

[krumme]

As Charlie at www.semiaccurate.com say, it will not be a barcelona nor a 4870 launch, http://semiaccurate.com/2011/09/07/bulldozer-finally-shipped-last-week/
There is a bug in the integer part, and the clock is not where they should be. The next stepping should be the one looking for. I think his remarks fit very well with the lack of rumours/BM. The performance is not there yet.
We have to remember that this part is for servers, its supposed to be moving those integers big time, and with a high clock. And when there is problems here, well, its not good.
I dont think this cpu was intended for the very small high-end desktop market. If video encoding is not your primary consern, just go and buy the sb. What BD will do is just force the price down a little.

 

[LOL_Wut_Axel]

As Charlie at www.semiaccurate.com say, it will not be a barcelona nor a 4870 launch, http://semiaccurate.com/2011/09/07/bulldozer-finally-shipped-last-week/
There is a bug in the integer part, and the clock is not where they should be. The next stepping should be the one looking for. I think his remarks fit very well with the lack of rumours/BM. The performance is not there yet.
We have to remember that this part is for servers, its supposed to be moving those integers big time, and with a high clock. And when there is problems here, well, its not good.
I dont think this cpu was intended for the very small high-end desktop market. If video encoding is not your primary consern, just go and buy the sb. What BD will do is just force the price down a little.

If there's problems with getting high integer performance, then why would they ship for revenue Interlagos already? I understand floating point performance matters more in server than desktop, but integer performance is still a huge factor for servers as well.