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

[BallaTheFeared]

7990 @ 1080p Max IQ.

 

[Gikaseixas]

Well, IDC is human like the rest of us and it's just annoying to see cringe worthy posts constantly on here; I will say he has a higher tolerance threshold for BS than most, and makes informative posts all the time.

The salient point being - many of the comments are originating from those who are making broad assumptions about how things are done from an engineering standpoint which are entirely nonfactual. It's not like intel has engineering team meetings that go like this: "Hey gets lets figure out how to SCREW the overclockers! MWAHAHAHHA!" "VICTORY ACHIEVED!" (High-fives all around at Haswell's release). Come on. This is not how things happen from an engineering standpoint. There is no conspiracy at play here.

I've said it before and i'll say it again - chips that overclocked as well as SB did come once in a blue moon, and chips that are worthy of a yearly upgrade also come only once in a blue moon. I've been in the game for a while and I seldom recall CPU releases so groundbreaking that it made me *want* to upgrade faster than my typical 2-3 year upgrade cycle. There have been a few outliers, but overall Haswell fits in very nicely. It still overclocks well and has a very nice IPC boost over IVB and SB - a 4.3ghz Haswell is roughly equal to a 5ghz SB. Actually, i've seen benchmarks with a 4.3ghz 4770k outperforming a 5ghz 2600k. The Haswell is a great CPU, especially for the mobile sector - I really feel like people have a combination of inflated expectations and rose-tinted glasses from the SB. Again, once in a blue moon.

Valid points, however that doesn't mean the engineering team is always spot on, they mess up sometimes. I just find it hilarious that the company gets defended at all costs by the same individuals and so on...
I'm waiting on the next step, i have faith that Intel will release something that we'll be able to get 4.8ghz out of it regularly. It has a decent IPC boost over IB but let's not pretend that by achieving 4.3ghz on Haswell (matching IB@4.8) is what we all expected. I expected to see Haswell do 4.8ghz and give me a motive to upgrade. Haswell @4.3 is a sidegrade.

 

[blackened23]

Valid points, however that doesn't mean the engineering team is always spot on, they mess up sometimes.

No disrespect intended, but it's also strange to suggest that intel somehow screwed up - Haswell was designed for mobility and extremely long battery life, they did not "mess up". Haswell may be the best and most successful chip that intel has ever released considering the current market landscape (eg. mobile). As well, Intel does not cater to extreme overclockers, it's not like their engineering efforts are fully devoted to that.

 

[Beavermatic]

Haswell was just one of those "sitting around twiddling our thumbs" releases from intel while they wait for AMD to catch up. Now that I think about it, Ivy Bridge probably was too.

 

[frozentundra123456]

No disrespect intended, but it's also strange to suggest that intel somehow screwed up; Haswell was designed for mobility and extremely long battery life, they did not "mess up". Haswell may be the best and most successful chip that intel has ever released considering the current market landscape (eg. mobile). As well, Intel does not cater to extreme overclockers, it's not like their engineering efforts are fully devoted to that.

We will have to wait and see, but the temperatures could be an issue in mobile. I know the heat output is lower, but temperature is a problem as well, since the CPU will throttle at a certain temperature no matter what the total heat output is.

Whether Intel "screwed up" or if this is just a basic problem of smaller dies, I don't know.

 

[Gikaseixas]

No disrespect intended, but it's also strange to suggest that intel somehow screwed up - Haswell was designed for mobility and extremely long battery life, they did not "mess up". Haswell may be the best and most successful chip that intel has ever released considering the current market landscape (eg. mobile). As well, Intel does not cater to extreme overclockers, it's not like their engineering efforts are fully devoted to that.

Best arch released by Intel? Don't agree, i think Conroe and Sandy are hard to top but that's just my opinion.

Since we're in a enthusiast forum (CPU and Overclocking) it's safe to assume that Haswell wasn't received very well. Intel might not cater to overclockers as you put it but when designing a K model what kind of buyers are they thinking of?

Again, in a overclocker perspective they messed up. Balla's the only user here who seems to be able to reach 4.8ghz.

 

[monstercameron]

Let me get this straight, they released a k model i7 only to not have it cater to enthusiasts and design for mobiles?

 

[BallaTheFeared]

Let me get this straight, they released a k model i7 only to not have it cater to enthusiasts and design for mobiles?

Yeah and they're going to put the same core design into billion dollar supercomputers, those bastards!

 

[guskline]

Gikaseixas: Excellent post. BTW, since your FX 8350 is already at 4.7Ghz if the FX 9590 is released (4.7Ghz base/ 5Ghz turbo) will you consider it?

What clock speed is your 8350 ?

I reach 4.6Ghz (below) with a 21 multiplier x 219 fsb combo. Much more stable at 1.46v than 23 x 200 and much cooler.

 

[monstercameron]

Yeah and they're going to put the same core design into billion dollar supercomputers, those bastards!

thats crazy bro, portable super computers on batteries with k-models, the future is now!

 

[Hulk]

If Intel has released Haswell at 4.8GHz and the performance gains over Ivy were 25% then people here would be crying foul because "they just raised the clockspeed my overclocked Ivy is just as good!"

 

[sushiwarrior]

If Intel has released Haswell at 4.8GHz and the performance gains over Ivy were 25% then people here would be crying foul because "they just raised the clockspeed my overclocked Ivy is just as good!"

But here AMD is releasing one model of chip, and not calling it a redesign or anything. Ivy/Haswell is not the same thing as Piledriver/Piledriver...

 

[Gikaseixas]

Gikaseixas: Excellent post. BTW, since your FX 8350 is already at 4.7Ghz if the FX 9590 is released (4.7Ghz base/ 5Ghz turbo) will you consider it?

What clock speed is your 8350 ?

I reach 4.6Ghz (below) with a 21 multiplier x 219 fsb combo. Much more stable at 1.46v than 23 x 200 and much cooler.

No i rather keep the 8350. I have my eyes on the i7 4770 I'm giving the i7 860 to a nephew

My 8350 is running @4.7 with a 21,5 multiplier X 218 fsb. Voltage is 1.48 and my H100i keeps it cool.

 

[guskline]

No i rather keep the 8350. I have my eyes on the i7 4770 I'm giving away my i7 860

My 8350 is running @4.7 with a 21,5 multiplier X 218 fsb. Voltage is 1.48 and my H100i keeps it cool.

Thanks for the info. Good luck with the 4770k.

 

[USER8000]

It will be interesting to see whether these chips are specially binned, and consume less power than a FX8320 or FX8350 at similar clockspeeds.

 

[Gikaseixas]

Thanks for the info. Good luck with the 4770k.

either that (given a new step is released) or i will go Intel Extreme.

 

[guskline]

either that (given a new step is released) or i will go Intel Extreme.

A person with a 2009 BMW X6 , 2012 Mitsubishi Pajero 3.8 V6 DESERVES a socket 2011 Intel Extreme

 

[Gikaseixas]

A person with a 2009 BMW X6 , 2012 Mitsubishi Pajero 3.8 V6 DESERVES a socket 2011 Intel Extreme

LOL i guess. Crazy about them trust me.
I really hope these FX's are 140w maximum and give us a reasonable bump in speed. Intel needs a wake up call, after all Haswell is a tock.

 

[Abwx]

4.6Ghz. 1.46v

@4.7 . Voltage 1.48

The new chips should run at those frequencies
with about 100mV lower voltage.

 

[galego]

SlowSpyder, excellent post. Particularly the pricing. I saw someone post the pricing at $200 for the top end FX 9590. If that's the case AMD will be back in the game. If the converse is true ($800), I doubt there will be many takers.

The $200 speculation can be found here

http://mygaming.co.za/news/hardware/55134-amd-fx-9590-and-fx-9370-processors-unveiled.html

and here

http://www.dailytech.com/GigaHertz+...s+First+50+GHz+FX+Processor/article31741c.htm

 

[guskline]

The $200 speculation can be found here

http://mygaming.co.za/news/hardware/55134-amd-fx-9590-and-fx-9370-processors-unveiled.html

and here

http://www.dailytech.com/GigaHertz+...s+First+50+GHz+FX+Processor/article31741c.htm

A big IF on price. I suspect that at 4.7/5 speeds AMD will require a water cooling system. Remember that AMD bundled both the 8150 and the 8350 with a water cooling system at a higher price. They would have to really drop the price of the current 8350 to sell the 9590 at $200. The $800 figure is most likely too high as the $200 figure is too low. My guess? @$300-350 to match the 3770k/4770k.

 

[galego]

Whats the source?

Anandtech: "We’ve asked AMD for details on the base clocks for the new parts, but so far have not yet received a response"

The source for the base clock of the 9590 is AMD

 

[frozentundra123456]

Do you have benchmarks for an 8350 running at 5ghz to confirm this statement, or are you just guessing? I seem to recall that the 8350, at stock, was only slightly behind Intel, sometimes even ahead in a few rare cases. You really don't think a 20% increase in clock speed will make a significant difference?




You aren't looking at the same image. The base clocks in the table referenced, which I was responding to, were listed as 3.7 or 4.0ghz, not 4.7.

Where did you hear the base clock would be 4.7ghz? That would just be silly. Why would a CPU with a max turbo of 4.7 also have a 4.7 base clock?

Looking at cinebench single threaded. 3770k is 66% faster, 4770k is 78% faster. The difference in max turbo for the new chip is +19%. Even assuming perfect scaling with clockspeed, it is not even close. There are a lot of other primarily single threaded benchmarks where the 3770k or 4770k are ahead by considerably more than 20%.

 

[guskline]

Silly or not, the report has the 9370 at 4.4base/4.7 turbo and the 9590 at 4.7 base and 5.0 turbo. I have my 8350 running 24/7/365 at 4.6Ghz (21 x 219) at a vcore of 1.462 max in the bios. I've bumped the vcore as high as 1.515 to boot at 5 Ghz so the figures being bantered around are clearly doable. I think there will not be much overclock headroom in the 9590 at 4.7/5 but it's likely AMD can release a stable chip at those speeds. The real questions will be power needed for vcore and heat output but again there is water cooling to address that. Perhaps "bragging rights" over Intel about having a 5 ghz processor might be a factor ( I know IBM has a 5.2Ghz processor but it's uber expensive). BTW my 3770k at 4.4Ghz beats my 8350 at 4.6Ghz in almost all benchmarks. The 8350 is a fast one though! Gikaseixas can chime in because his 8350 is at 4.7ghz which would be the base for the higher end 9590.

 

[Erenhardt]

Looking at cinebench single threaded. 3770k is 66% faster, 4770k is 78% faster. The difference in max turbo for the new chip is +19%. Even assuming perfect scaling with clockspeed, it is not even close. There are a lot of other primarily single threaded benchmarks where the 3770k or 4770k are ahead by considerably more than 20%.

Yes, that is why we don't look at those benchmarks. It is jaguar thread all over again...