bit-tech.net

AMD FX-8120 review

Comments 26 to 50 of 59

Reply
Harlequin 27th July 2012, 19:53 Quote
so if AMD goes bankrupt , how wouldnt it happen?
.//TuNdRa 27th July 2012, 20:10 Quote
If AMD goes Bankrupt; Intel wouldn't be allowed to buy them up in Liquidation, another company would have to take possession, to ensure that Intel still has competition.
Elton 27th July 2012, 20:11 Quote
You know what I realized? It seems that Bulldozer was basically an old plan drawn up during the Pentium 4 days. It has all of the hallmarks of it. A focus on parallelism, an emphasis on a form of hyperthreading, a change in cache architecture, a reliance on threading rather than a stronger FPU.. Hell even the extended pipeline cycles is reminiscent of Pentium 4.

What could save AMD though is if they manage to combine Bulldozer and their K8 architecture (just like Intel did with Nehalem (Pentium 4 + Core 2 Duo)). They managed to drag the K8 into K10 and compete with the height of Core 2 Duo. If they could leverage that with the innovations they made with Bulldozer and say, increase the efficiency of the memory controller and integer units and we could have on our hands something innovative and blazing fast from AMD.

That is of course if AMD goes through with it. And if they'll go back to hand designing things..
Harlequin 27th July 2012, 20:26 Quote
Quote:
Originally Posted by .//TuNdRa
If AMD goes Bankrupt; Intel wouldn't be allowed to buy them up in Liquidation, another company would have to take possession, to ensure that Intel still has competition.

and what if no one bought them?
Elton 27th July 2012, 20:50 Quote
Quote:
Originally Posted by Harlequin
and what if no one bought them?

I'm pretty sure IBM, Fujitsu or some other guys would snatch it up. The problem though is that Intel is one of the few Fabbed companies.
Action_Parsnip 27th July 2012, 20:58 Quote
Quote:
Originally Posted by .//TuNdRa
I was suspecting that was the case, from the absolute lack of information i'd heard about a different stepping, while AMD was busy pushing Piledriver as hard as they could.

It was supposed to happen but with Vishera 2 months away the chances of a B3 are nill.

'bdver 1' (the whole arch is I believe called bdver) is not quite as bad as the bit-tech test suite makes out, but still bad.

The bit-tech cpu test suite is quite awful.

Piledriver is bdver 2, but the really interesting stuff will be bdver 3 & 4.
xxxsonic1971 27th July 2012, 23:18 Quote
i'll be keeping my i7 920 for a while.
rollo 27th July 2012, 23:55 Quote
i7 920 still kicking it in amds face what is it now 3 years later lol and thats the bottom end chip in the series,
Amd are truly not competing with intel anymore and if they dont start to improve they could be outta business. which is bad for everyone, gone are the days when they did produce fast chips
dancingbear84 27th July 2012, 23:57 Quote
I love AMD. I now rock on a maximus v gene and a 2500k
Anfield 28th July 2012, 00:00 Quote
Quote:
Originally Posted by Harlequin
so if AMD goes bankrupt , how wouldnt it happen?

As the financial crisis has taught us, they'll simply get tax money if they are risk of going under, so no need anymore for buying amd cpus as a charitable act.
fluxtatic 28th July 2012, 05:43 Quote
Quote:
Originally Posted by Anfield
As the financial crisis has taught us, they'll simply get tax money if they are risk of going under, so no need anymore for buying amd cpus as a charitable act.

Nah, that's just for banks and big manufacturers. AMD is fabless, so they can go right the hell under and the government won't care.

They're being propped up now by their low- and mid-range. Hard to tell, since enthusiast sites don't exactly cover the FM1 platform a whole lot, but that's where AMD makes money on the processor side these days.

They've sunk enough into it now that they'll keep on with the BD-derivations to the end of the roadmap they laid out (2014 or so, iirc), I suspect. I wouldn't be surprised if they bailed after that and concentrate on the low- and mid-range in the consumer space. What'll really kill them is if they can't come up with competitive server procs. AMD and Intel both make an absolute killing on server platforms, and things aren't so rosy for AMD there, either. BD is awesome at some very specific server applications, but that can't carry them. They need to get back to the drawing board and figure out how they'll beat Intel in the server space again.

For myself, I'll likely build my wife a PC on the FM1 platform as my next build. When it comes time to replace my own desktop (a ways off yet - my PII X3 is still going strong) I get the feeling I'll switch teams. I don't particularly want to, but when my old MB died, I got a AM3+ board, wanting to be ready for the future. Imagine my disappointment when the procs finally dropped and I realized there was no reason to replace a processor I paid $75 for with one that cost double to triple.
Elton 28th July 2012, 05:56 Quote
For some reason I get the inkling of a feeling that AMD might end up combining K10 and BD architectures. I'm sure there's still some way to pull it off. After all Intel did it before.
Action_Parsnip 28th July 2012, 13:14 Quote
Quote:
Originally Posted by fluxtatic
They've sunk enough into it now that they'll keep on with the BD-derivations to the end of the roadmap they laid out (2014 or so, iirc), I suspect. I wouldn't be surprised if they bailed after that and concentrate on the low- and mid-range in the consumer space. What'll really kill them is if they can't come up with competitive server procs. AMD and Intel both make an absolute killing on server platforms, and things aren't so rosy for AMD there, either. BD is awesome at some very specific server applications, but that can't carry them. They need to get back to the drawing board and figure out how they'll beat Intel in the server space again.

No they'll keep BD going until the end of time. K10 is dead and buried. Modernising K10 and 'fixing' BD would take up similar development budgets. Don't forget BD has only been 'warmed over' once and made into Trinity and looks pretty good for it. I look on BD as a rough CPU equivalent to the 2900xt, the core idea (VLIW5) is sound but substantial chunks of the design are inadequate and need rebuilding. The process technology, like for the 2900xt, also looks (and since Trinity, *looked)....'iffy'.

AMD have commited to 'warming over' BD every 12 months. Version 2 for desktops and servers (Trinity & Vishera/Abu Dhabi) is inbound shortly and looks to be substantially ...'less bad'. BD never struck me as being a total lost cause and has a very good chance of eventually maturing into something really rather good. Any serious, grown-up look at where things have gone wrong with BD says that it is let down (seriously) by deficiencies in a few key areas. The philosophy of the design and it's basic topology are not inherently wrong or misguided. David Kanter, the owner of realworldtech.com (excellent source of very indepth articles) basically surmised that it appears some areas of the architecture recieved a bare minimum of development time whereas others were very accomplished in their implementation.

Any investigation of BD worth a damn has flagged up 3 areas of concern:

1). Some serious L1 cache thrashing.
2.) Branch missprediction penalties are very, very high (branch predicition is however very good)
3.) Outside of server applications, the L2 cache is far too slow and a big drag on performance.

Piledriver nibbles away at the process problems and a host of small architecture niggles to make a much improved whole but addressing the issues listed above, especially points 1) and 2) will have to wait for the third and fourth revision of BD.
Quote:
Originally Posted by Elton
For some reason I get the inkling of a feeling that AMD might end up combining K10 and BD architectures. I'm sure there's still some way to pull it off. After all Intel did it before.

Intel never merged Pentium M (Conroe predecessor) and Pentium 4. Hyper Threading and a trace cache are not things that made the P4 a P4, a very long pipeline, high latencies for even common, simple instructions and horrendous branch-missprediction penalties were those things.
Harlequin 28th July 2012, 13:20 Quote
http://www.anandtech.com/show/5057/the-bulldozer-aftermath-delving-even-deeper/


^^ a good read on the opinion of where bulldozer went wrong and actually where it went right
Elton 28th July 2012, 21:05 Quote
Quote:
Originally Posted by Action_Parsnip

Intel never merged Pentium M (Conroe predecessor) and Pentium 4. Hyper Threading and a trace cache are not things that made the P4 a P4, a very long pipeline, high latencies for even common, simple instructions and horrendous branch-missprediction penalties were those things.

Fair enough. Although the trace cache and Hyperthreading were things introduced by Pentium 4. The missteps were byproucts of trying to fit in too much too early.

Mind you you still see traces of Pentium 4 and Pentium M in Nehalem and Sandybridge. Albeit without the deficiencies of what you stated.

As for Bulldozer right now, I'm inclined to agree with you, BD needs quite a bit of refinement and in maybe a weird world where it would be easy to combine technologies but I guess i'm over-simplifying it. You've pretty much pointed out BD's weaknesses, the only problem now is how will AMD address the misprediction penalty?
.//TuNdRa 28th July 2012, 22:15 Quote
Who knows? At this point in time; they're just upping clock speeds and IPC, which means they'll have to alleviate some of that branch prediction penalty that's kicking the IPC in the teeth.

Does mean, however, that Piledriver might have lower temps. Bulldozer is wieeeerd with the temperature scales. Phenom II Chips went through reasonable curves. BD chips just jump. There's no curve there. 30-*load*-50. 51. 52. So on and so forth. Even did the same thing under water, cheeky bugger! (Yes. I had a good application of thermal paste. Yes. The block was unclogged. And yes, I checked all of this on first install on air anyhow. BD is just weird.)
Elton 28th July 2012, 22:28 Quote
Quote:
Originally Posted by .//TuNdRa
Who knows? At this point in time; they're just upping clock speeds and IPC, which means they'll have to alleviate some of that branch prediction penalty that's kicking the IPC in the teeth.

Does mean, however, that Piledriver might have lower temps. Bulldozer is wieeeerd with the temperature scales. Phenom II Chips went through reasonable curves. BD chips just jump. There's no curve there. 30-*load*-50. 51. 52. So on and so forth. Even did the same thing under water, cheeky bugger! (Yes. I had a good application of thermal paste. Yes. The block was unclogged. And yes, I checked all of this on first install on air anyhow. BD is just weird.)

Ever OC'ed Pentium 4? While it was hot from the beginning, the leakage on some chips that occured at around 3.9+GHz was so obscene it beggared belief. Although I will admit, BD on load reveals an inherent issue since it's idle throttling is mighty impressive. Remember that this chip consumes a boatload of power. Couple it with an architecture that was desgined to run at higher clocks (Pentium 4 eh?, okay not really) but didn't make it to those clocks and you have some pretty crazy leakage.
.//TuNdRa 28th July 2012, 22:55 Quote
Yup. That said; I can push this chip to well over the 4.65ghz that they achieved, I could even hit 5ghz if I balanced the Vcore enough, but it just gets so damn hot.

I would like to see the exact bios setttings used while benching, though. I'm curious. AMD have a load of safeguards in the Bulldozer chips that you have to disable to get them to move anywhere near quickly, from Application Power Management, which is a plug-in for turbo, which means it'll actually friggin' downclock if too many cores are loaded, to the typical Hardware Thermal Control, which just insta-throttles if the CPU even gets near 50 degrees centigrade. Which it will.

Also pushing the motherboard can yield better performance. The HTT link is plugged into the Memory Controller, so the higher you can get the motherboard and northbridge; the more memory bandwidth you open up for BD to work with, which can really boost scores in some tests.
Elton 28th July 2012, 23:25 Quote
I gotta take down notes for when I get myself an AMD build. :D I think it's running into the same frequency scaling issues that all chips ran into. With the higher frequency leads to higher voltages and leakage, and with that comes increased heat. But if anything, there is some truth and yet still some falsehoods about GloFlo's transistor manufacturing process. On one hand it's made to spec. On the other, the spec has dramatically upscaling leakage.
Harlequin 28th July 2012, 23:26 Quote
i cant see Piledriver being a huge step from BD - its the one after which should fix the issues.
Elton 28th July 2012, 23:40 Quote
Quote:
Originally Posted by Harlequin
i cant see Piledriver being a huge step from BD - its the one after which should fix the issues.

It isn't, it's claimed to have 7-15% better performance. But that's to be expected. It's still using the same node, the same general architecture. The issues can only be incrementally concluded anyways.
Isitari 29th July 2012, 20:20 Quote
Quote:
Originally Posted by Harlequin
and what if no one bought them?

A license to produce X86 is worth its weight in gold if done right.
Harlequin 29th July 2012, 20:30 Quote
any company would have to start from the ground up
.//TuNdRa 29th July 2012, 21:47 Quote
Not really. AMD has a lot of expertise within them, it's just a long complicated chain of decisions that's lead them down this path. Once they're on the path; they need to continue. You cannot afford, in this line of business, to turn around and change your mind unless the resultant item is so ungodly awful that it's worth more if you go "Screw that. We need something brand new."

Bulldozer isn't that bad. It's slow, it's hot, but it has the ability to evolve.
ch424 29th July 2012, 21:55 Quote
Quote:
Originally Posted by .//TuNdRa
It's not the engineers. AMD is using artificially generated CPU Silicon designs, not the original Hand-designed, carefully planned pathways that wound up more efficient and smaller. Which is letting them down in some areas.

This isn't true - AMD's FPU shrank area by 30% and reduced critical path by 20% when they switched to using synthesis tools.
Log in

You are not logged in, please login with your forum account below. If you don't already have an account please register to start contributing.



Discuss in the forums