AMD unleashes first ever commercial “5GHz” CPU, the FX-9590

[dastral]

Sorry, but the 8 and 9 series cpus have 8 true cores. You must be confounding it with Intel chips like the i7-3770k which has 4 real cores plus 4 virtual cores (the virtual cores are not "real")

This is joke right ? You CAN NOT be serious posting this.
While i agree AMD's "bonus cores" are better than Intel's "bonus cores".
Saying 8000 and 9000 have "8 true cores" is just a ignorance, or even worse a blatant lie/manipulation.

Assuming the price is right (let's not get into TDP) there is probably a small market for this kind of CPU :
As fast as a 4770 in single thread, and as good as a 3930K in multi-thread.

 

[inf64]

8T PD/BD design has 8 true integer cores and 4 FP "cores" that can run 2 treads each. So in any case except AVX256bit you have 8T execution capability, just like with i7. That's why performance is on par with i7 in MTed workloads- AMD's "cores" scale much better.

 

[guskline]

Nice.
+1 to Povray multithreaded benchmark.
If you can, Thanks.
http://www.povray.org/beta/

I'll be glad to run it this weekend. It was early this morning @1 AM eastern time when I finally got the chart posted and downloaded Povray 3.7. It takes time to run (actually have to run a single thread then 8 threads for both the 3770k and 8350 stock and OC'd) so that will take time but I'll do my best to get it done this weekend.

As to the controversey and acrimonious tone on this thread I'm thinking why complain about FX 9590 power draw when in the GPU area we are running Titans, 7970s soon to be 7990 and so on.

Moreover, if I'm AMD I have to bring this out to keep some semblance of hope on the high-end desktop arena. Sure Haswell has its critics about not being the greatest OCer etc but Intel is still getting ink about new desktop cpus and where is AMD getting ink from???

From a practical end these releases(9370/9590) show the PileDriver still has life in it and apparently AMD has found a way to produce 4.7/5.0Ghz chips that are stable. It helps keep enthusiasts in the AMD fold (supposedly) without switching over to Intel and Haswell. My testing, abeit on a very limited amount of benchmarks, shows that a 4.7Ghz PileDriver narrows the gap or beats a stock 3770k. I'm unsure about a 4770k. I won't argue the power draw debate, Intel wins. Seriously, though, if you are running Titans in SLI is power draw as much of a priority as say brute cpu power? I think AMD is going to milk this area as long as they can and I don't blame them. Hopefully, they have some R&D $$ left to perfect and release a SteamRoller cpu that puts serious pressure on Intel both from a raw performance position AND a powerdraw position. Right now AMD's only hope on the desktop high end is the BRUTE strenghth of a heavily OC'd PileDriver.

Just my thoughts.

 

[inf64]

Good post guskline, you nailed it

 

[guskline]

BallaTheFeared: Is that 8.33pt figure in Cinebench 11.5 for a stock 4670k or is that OCed?

 

[BallaTheFeared]

BallaTheFeared: Is that 8.33pt figure in Cinebench 11.5 for a stock 4670k or is that OCed?

It's at 4.8GHz.


Stock is 6.something.

 

[Abwx]

While i agree AMD's "bonus cores" are better than Intel's "bonus cores".
Saying 8000 and 9000 have "8 true cores" is just a ignorance, or even worse a blatant lie/manipulation.

It scale way more than 4X when going from single thread
to 8 threads , how could a 4 cores scale more than 4X..?.

http://techreport.com/review/23750/amd-fx-8350-processor-reviewed/12

 

[Obsoleet]

Nice! I've been waiting for 5ghz forever. Yeah, I know it doesnt really mean anything but I don't care. All CPUs are fast enough (my current favorite is the A10 6800K and it has enough CPU crunching power for me, but the APU features I want).

5GHZ sounds great to me.

 

[Vesku]

This is joke right ? You CAN NOT be serious posting this.
While i agree AMD's "bonus cores" are better than Intel's "bonus cores".
Saying 8000 and 9000 have "8 true cores" is just a ignorance, or even worse a blatant lie/manipulation.

Assuming the price is right (let's not get into TDP) there is probably a small market for this kind of CPU :
As fast as a 4770 in single thread, and as good as a 3930K in multi-thread.

The Pentium 4 EE sold so this probably will too. Interestingly they mainly went into the same systems, enthusiast prebuilt.

 

[SlowSpyder]

So I was looking over at this post in the Intel roadmap thread. The i7-4820K looks to be a 4770K with 2MB more cache, a 200Mhz higher base clock (same 3.9GHz turbo), more memory controllers, and no IGP. TDP is up 55% to 130 watts. There is a number of posters from this thread I expect in that thread expressing outrage, but I'm not seeing it yet. Wierd.

 

[inf64]

So I was looking over at this post in the Intel roadmap thread. The i7-4820 looks to be a 4770K with 2MB more cache, a 200Mhz higher boost clock, more memory controllers, and no IGP. TDP is up a 55% to 130 watts. There is a number of posters from this thread I expect in that thread expressing outrage, but I'm not seeing it yet. Wierd.

4820 is IB-E based . Not Haswell. But it will probably end up performing the same as 4770K while having 130W TDP spec. I doubt any of the posters who complain today about new high perf./high TDP FX will complain about 4820.

 

[Mallibu]

Sorry, but the 8 and 9 series cpus have 8 true cores. You must be confounding it with Intel chips like the i7-3770k which has 4 real cores plus 4 virtual cores (the virtual cores are not "real").

I said that Crysis 3 is optimized for four cores/threads, not that the engine cannot use more cores/threads. At very high quality settings, Crysis 3 loads a 4-core chip above the 95%, but fails to max. load 6 and 8-core chips



As said in #232 the top Centurion chip will run Crysis 3 faster than the i7-3930K and i7-3970x.



You omit to mention that the claim was made in a very specific scenario: it was about using the 8-core chip in multithreaded scenarios not in games developed for single or 3-cores.

In multithreaded scenarios the i7-3770k (HT activated) can be up to a 42% slower than the FX-8350. Imagine how much slower will be the 2500k... No wait, you don't need to imagine it

http://openbenchmarking.org/embed.php?i=1210227-RA-AMDFX835085&sha=293f200&p=2

2500k: 36.44
3700k: 33.05
8350: 23.34

The 2500k is a 56% slower. Therein the claim made by another poster who correctly said (bold from mine):


Moreover, the 8350 @ 4.6GHz scores 20.30. The FX-9590 would break the 20 seconds barrier with easiness.


Eurogamer discussed precisely this and did a poll with interesting results:


http://www.eurogamer.net/articles/digitalfoundry-future-proofing-your-pc-for-next-gen

Actually the I5s & I7s are faster in everything, but keep living in your dream world, where repeating the Open benchmark from Linux, and the article from developers, a dozen times per day, will actually make people believe AMD cpus are better.
In reality though, AMD is slower in 95% of the situations, consumes double the power, and has 20% market share as a consequence. Facts.

 

[SlowSpyder]

4820 is IB-E based . Not Haswell. But it will probably end up performing the same as 4770K while having 130W TDP spec. I doubt any of the posters who complain today about new high perf./high TDP FX will complain about 4820.

Ah yea, I forgot it is Ivy. But I think my point still stands, as you said.

 

[BallaTheFeared]

So I was looking over at this post in the Intel roadmap thread. The i7-4820K looks to be a 4770K with 2MB more cache, a 200Mhz higher base clock (same 3.9GHz turbo), more memory controllers, and no IGP. TDP is up 55% to 130 watts. There is a number of posters from this thread I expect in that thread expressing outrage, but I'm not seeing it yet. Wierd.

TDP isn't actually the concern, it's how high does it go beyond that number that is troubling.

Well that and using more power while being slower, but that's something else.

TDP in of itself is nearly worthless, much like the i7 and i5 share the same TDP, but the i7 uses more power.

 

[SlowSpyder]

TDP isn't actually the concern, it's how high does it go beyond that number that is troubling.

Well that and using more power while being slower, but that's something else.

TDP in of itself is nearly worthless, much like the i7 and i5 share the same TDP, but the i7 uses more power.

I get a 5.7% higher base clock than the 4770K in an architecture that is some ~4-5% slower than Haswell (same 3.9GHz turbo). And, a 55% higher TDP. Both at Intel 22nm, right? So it stands to reason as you overclock/add voltage both will go above and beyond their rated TDP at a similar rate, wouldn't they?

 

[guskline]

It's at 4.8GHz.


Stock is 6.something.

Thanks. Still impressive for a 4 core with no hyperthreading. I think Intel has just about gotten all it can out of these cores. I'm not advocating "moar cores" but I am extremely impressed with the combination of high cpu power and low power consumption Intel has achieved. Amazing.

 

[SlowSpyder]

Thanks. Still impressive for a 4 core with no hyperthreading. I think Intel has just about gotten all it can out of these cores. I'm not advocating "moar cores" but I am extremely impressed with the combination of high cpu power and low power consumption Intel has achieved. Amazing.

I didn't realize the 4670K didn't have HT. Balla, that is a pretty damn good score for a true quad! My 1090T @ 4.03GHz/2.6Ghz get ~7.1 (I'd have to run it again to know for sure). An that's with two more true cores at work...!

 

[Vesku]

So I was looking over at this post in the Intel roadmap thread. The i7-4820K looks to be a 4770K with 2MB more cache, a 200Mhz higher base clock (same 3.9GHz turbo), more memory controllers, and no IGP. TDP is up 55% to 130 watts. There is a number of posters from this thread I expect in that thread expressing outrage, but I'm not seeing it yet. Wierd.

I haven't seen much praise for its predecessor the 3820, it's not a great CPU choice really. If you are going for the 2011 platform it's kind of silly to get a 4 core instead of a 6 core, unless you very specifically need the increased memory bandwidth over 1155/1150 platforms.

But it's still not as much of an oddball as Pentium 4 EE and FX 9xxx.

 

[sushiwarrior]

I think we can all safely agree that it's better AMD is releasing SOMETHING at least, right? It's power hungry, it could be expensive, but hey, it's better than giving up completely.

 

[BallaTheFeared]

I didn't realize the 4670K didn't have HT. Balla, that is a pretty damn good score for a true quad! My 1090T @ 4.03GHz/2.6Ghz get ~7.1 (I'd have to run it again to know for sure). An that's with two more true cores at work...!

It's as fast as my i5-2500k @ 5.3GHz with faster ram in cinebench, and that required a decent custom water loop to pull off.

My 1090T at 4.5/3.2 got 7.87

 

[Vesku]

I think we can all safely agree that it's better AMD is releasing SOMETHING at least, right? It's power hungry, it could be expensive, but hey, it's better than giving up completely.

I'd rather they had taken the money spent on releasing this to improve product releases relating to GPUs and Steamroller. If anything this is just a sign that the pattern of new product delays AMD has is not going to be broken anytime soon. Why release these if they will have steamroller, even if it's just Kaveri APUs, out before December?

 

[mrmt]

I think we can all safely agree that it's better AMD is releasing SOMETHING at least, right? It's power hungry, it could be expensive, but hey, it's better than giving up completely.

8C FX is a niche, 8C FX 220W is a niche inside a niche. I don't think it does make any difference in AMD situation against Intel, much less in their revenue line. It would be a nice product if Intel didn't have LGA2011 chips, because once you factor them, 220W FX loses whatever Halo effect they might have.

IMO this has more to do with AMD trying to coerce MB makers to field 220W+ designs than anything else.

 

[mrmt]

I'd rather they had taken the money spent on releasing this to improve product releases relating to GPUs and Steamroller. If anything this is just a sign that the pattern of new product delays AMD has is not going to be broken anytime soon. Why release these if they will have steamroller, even if it's just Kaveri APUs, out before December?

What if Steamroller doesn't give as much improvements as everyone is expecting?

 

[ShintaiDK]

I think we can all safely agree that it's better AMD is releasing SOMETHING at least, right? It's power hungry, it could be expensive, but hey, it's better than giving up completely.

To release a new SKU with 220W(+) TDP that needs new motherboards, new validation and so on. Just to sell what, maybe 1000 chips? Its simply terrible financially. Even the FX8xxx series cant substain itself with is very low sales volume.

Then you could say it got PR value. But its rather a backfire in the quite large scale.

 

[monstercameron]

I'd rather they had taken the money spent on releasing this to improve product releases relating to GPUs and Steamroller. If anything this is just a sign that the pattern of new product delays AMD has is not going to be broken anytime soon. Why release these if they will have steamroller, even if it's just Kaveri APUs, out before December?

seems to be amds way, the phenom II line with turbo, brazos 2.0, richland and centurion. all with better clocks and turbos. besides the extra 10 seconds it takes to finish encoding or the 10 fewer fps playing a game a 720 on low while using a titan wont kill you to wait until Q4