| Wednesday, August 27 2008 |
The transition of the K10 architecture by AMD to the 45nm silicon fabrication process is stirring up interesting revelations these days. First, it was about surprisingly low power consumption of the quad-core Phenom parts, and then about the overclocking headroom those 45nm parts provided, at least the engineering samples did so far. And now, news coming in that AMD could be resurrecting the "FX" series of extreme performance products. Over the past three or so years, the performance trail AMD products had over Intel's made it close to impossible for AMD to sell parts that provide performance tuning advantages such as unlocked FSB multiplier settings for a premium, like it did back when K8 reigned the performance segment. "Black Edition" chips made up for that deficit by providing consumers overclocking advantages while not charging a significant premium and at the same time, safeguarding the "FX" title, not letting it dilute.
Come AMD Deneb core and lot seems to be on offer. To begin with, unlike the Windsor core that had a maximum FSB multiplier of 16.0x, initial reports suggest the Deneb to sport a maximum 25.0x multiplier, 200 MHz x 25 = 5.00 GHz, with the FSB left to play with. Considering at 2.30 GHz the Deneb draws in 57.3 W (according to findings), it should still leave enough room for AMD to sell premium products clocked at high frequencies.
From Reviewage's findings, there seem to be two Phenom FX processors in the making. The numbering seems to take off where it last left at the Athlon64 FX 74. The two chips, Phenom FX 80 and Phenom FX 82 could be clocked at 4.00 GHz and 4.40 GHz respectively (stock speeds). An interesting statement is that at 4.00 GHz, the Phenom FX 80 should outperform an Intel Kentsfield core clocked at 5.00 GHz, implies it has to be faster than the Kentsfield on a clock-to-clock basis. This opens up an interesting debate on how these parts compare to the succeeding Yorkfield chips. This should also open gates for several models to enter the market at various clock speeds.
Source: Reviewage
Come AMD Deneb core and lot seems to be on offer. To begin with, unlike the Windsor core that had a maximum FSB multiplier of 16.0x, initial reports suggest the Deneb to sport a maximum 25.0x multiplier, 200 MHz x 25 = 5.00 GHz, with the FSB left to play with. Considering at 2.30 GHz the Deneb draws in 57.3 W (according to findings), it should still leave enough room for AMD to sell premium products clocked at high frequencies.
From Reviewage's findings, there seem to be two Phenom FX processors in the making. The numbering seems to take off where it last left at the Athlon64 FX 74. The two chips, Phenom FX 80 and Phenom FX 82 could be clocked at 4.00 GHz and 4.40 GHz respectively (stock speeds). An interesting statement is that at 4.00 GHz, the Phenom FX 80 should outperform an Intel Kentsfield core clocked at 5.00 GHz, implies it has to be faster than the Kentsfield on a clock-to-clock basis. This opens up an interesting debate on how these parts compare to the succeeding Yorkfield chips. This should also open gates for several models to enter the market at various clock speeds.
Source: Reviewage
User comments
the Kentsfield is also higher clocked, and overclocks almost 100%.. :
Here, have a read of this, it will explain it better than I can...............
http://www.tek-tips.com/faqs.cfm?fid=4102
http://www.tek-tips.com/faqs.cfm?fid=4102
by: Bluefox1115I meant at the same clock speeds.
the Kentsfield is also higher clocked, and overclocks almost 100%.. :
I can honestly see an amd quad core 45nm at 2.8 to 3.0 ghz stock running in the 30 to 40 degree mark on idle. and at 3.5? 50C tops with decent cooling. isnt amd updating the coolers as well, for the new chips? they are going to be more effective and not just a hunk of aluminum,?
yea. Intel has been shown to beat AMD at stock speeds as well.. about 200 to 400mhz wise in actual speed difference.
by: Tatty_OneWell im sure they would, and funny enough they are goin to put a increase of cache on the CPU's, but why haven't they done this earlier? because its not that easy to do, and its very expensive to manufacture. Intel can do this and not even sneeze at it, big woop if it doesn't work as good, not like its going to hit there pockets hard, but if it does to AMD its not good at all, thats alot of money to them to waste if it doesn't do the job.
Do you not think that AMD would bolt on extra L2 memory if it would bolster their performance to any degree?.....of course they would, it would pay for itself in a month because so many more people would buy AMD if they were competative....no, as I said, the architecture, unlike Intels would benefit little, AMD can swop out the L2 so much quicker than Intel so it dont need the huge amount's Intel does.
by: Tatty_OneI understand what you are trying to say.. L2 cache has much better latency than L3 cache, and L1 cache has even lower latency than L2.. lower latency= higher performance. Much like RAM, 4-4-4-12 cas has lower latency and response time than 5-5-5-15.. and althought I understand your point, that post is way outdated and multiple cores with multiple caches were not taken into consideration, I would like to see an updated version of that thesis though, it's interesting. :)
Here, have a read of this, it will explain it better than I can...............
http://www.tek-tips.com/faqs.cfm?fid=4102
What Tatty is trying to say is that without a huge L2 cache that intel chips can't perform well whereas with AMD that much cache isn't needed because AMD had worked out a better bus than intel.
that is also true. no actual FSB, just a reference clock, and onboard memory controller.
The original subject of this thread is by far the most ridiculous pile of BS I've seen on tech forums...
K10.5 v K10
more transistors
more pins
65nm v 45nm*
PCIe 3.0 v 2.0
HyperTransport 3.1 v HyperTransport 1/2
uses less power to run
outputs less heat
better performance
ddr3 v ddr2
much much better platform (motherboards, chipsets & SB800 coming 1/2009)
6mb L3 cache v 0mb L3 cache
2.3GHz Deneb consumes only 57.3W v 2.3ghz phenom 120/157W (4ghz = 120w approx)
Immersion lithography
4th generation of strained-silicon
new leading edge technologies??????
used low or High-K metal gate technology
This is not the DENEB FX Series they are talking about but the 3+GHz Denebs for OEM and average users.
http://news.softpedia.com/news/AMD-039-s-45-Nanometer-Deneb-Core-Beats-3-2-GHz-83973.shtml
AMD's 2007 analyst day. what a change (to counter) and hopefully better to come
http://techreport.com/articles.x/13792/1
more transistors
more pins
65nm v 45nm*
PCIe 3.0 v 2.0
HyperTransport 3.1 v HyperTransport 1/2
uses less power to run
outputs less heat
better performance
ddr3 v ddr2
much much better platform (motherboards, chipsets & SB800 coming 1/2009)
6mb L3 cache v 0mb L3 cache
2.3GHz Deneb consumes only 57.3W v 2.3ghz phenom 120/157W (4ghz = 120w approx)
Immersion lithography
4th generation of strained-silicon
new leading edge technologies??????
used low or High-K metal gate technology
This is not the DENEB FX Series they are talking about but the 3+GHz Denebs for OEM and average users.
http://news.softpedia.com/news/AMD-039-s-45-Nanometer-Deneb-Core-Beats-3-2-GHz-83973.shtml
AMD's 2007 analyst day. what a change (to counter) and hopefully better to come
http://techreport.com/articles.x/13792/1
No high k, nothing fancy used, get used to it that phenom is spent as far as clock rate goes.
by: X1REMEBullshit :)
K10.5 v K10
6mb L3 cache v 0mb L3 cache
2.3GHz Deneb consumes only 57.3W v 2.3ghz phenom 120/157W (4ghz = 120w approx)
K10 = 2MB L3 Cache
2,4Ghz Phenom 9750 = TDP 95W
Deneb 2,3Ghz TDP ? no one know it.;)
by: TheGuruStud
No high k, nothing fancy used, get used to it that phenom is spent as far as clock rate goes.
nothing fancy used. well i think 45nm is fancy enough for amd not mentioning plenty of tweaks as already said
actually nobody really knows if high k is used or not which is why i said low/high k
by: BvB123
Bullshit :)
K10 = 2MB L3 Cache
2,4Ghz Phenom 9750 = TDP 95W
Deneb 2,3Ghz TDP ? no one know it.;)
my mistake, yes 2mb l3 cache and as for your second, actually a few sites mention this including Chinese sites who have demoed it, and if you can be bothered to look its still there and if it was fake am sure as you know amd would have pulled it from the ears.
by: X1REME*sighs* 45 nm is not a clock increasing god. We're limited by the physical properties of the materials used and the exact architecture itself (K8 hello, you can make K8 smaller all you want, but it won't clock hardly any higher).
nothing fancy used. well i think 45nm is fancy enough for amd not mentioning plenty of tweaks as already said
by: TheGuruStudactually what you will get is more transistors which makes all the difference e.g power - higher clock (R6 vs R7 anybody)
*sighs* 45 nm is not a clock increasing god. We're limited by the physical properties of the materials used and the exact architecture itself (K8 hello, you can make K8 smaller all you want, but it won't clock hardly any higher).
by: X1REMEWhich is not what deneb is unless you're specifically talking cache, but we're not. It's about clock rate and phenom doesn't have it. End of story.
actually what you will get is more transistors which makes all the difference e.g power - higher clock (R6 vs R7 anybody)
the acc+ will work even better in sb800 chipsets, since amd is working towards a spider (own platform) all these little tweaks will have a massive performance boost (ht 3.1-pcie3-ddr3-sb8xx-etc)
Hold your horses! Please, wait until actual launch of Denebs takes place. We will wait and see what happens then... :rolleyes:
by: TheGuruStud""It's about clock rate"" by reducing to 45nm, you will get much more transistors (needs less power), more pins, more cache, better heat rate etc = higher clocks:toast:
Which is not what deneb is unless you're specifically talking cache, but we're not. It's about clock rate and phenom doesn't have it. End of story.
and again.... back to my earlier post about the gap from 3.0GHz to 5GHz...
rev. D architecture (definitely will come earlier if they are to compete with the mighty i7) :nutkick:
http://news.softpedia.com/news/AMD-Unleashes-Hydra-8-Core-Competition-for-Nehalems-84982.shtml
what if these are the deneb fx (keep dreaming)
http://news.softpedia.com/news/AMD-Unleashes-Hydra-8-Core-Competition-for-Nehalems-84982.shtml
what if these are the deneb fx (keep dreaming)
my point: there is a few sites i checked which state the k10 is very adaptable including amd`s site and by the looks of it, they will stick to it through 2009.
AMD said they are even thinking of changing the name to k11 (which is k10/k10.5/K10.5 rev. D)
AMD said they are even thinking of changing the name to k11 (which is k10/k10.5/K10.5 rev. D)
by: X1REMEHT3.1 isn't going to do squat for performance. The current HT isn't even close to maxed out, neither is PCIe2. Red Herrings to take away from the fact that nothing significant has changed in the architecture besides the die shrink.
the acc+ will work even better in sb800 chipsets, since amd is working towards a spider (own platform) all these little tweaks will have a massive performance boost (ht 3.1-pcie3-ddr3-sb8xx-etc)
And even if 45nm do OC to 4GHz and beyond, Phenom is still slower clock for clock vs current Intel Quads. Tweaks and Die shrinks aren't going to change that fact. They need a whole new architecture to pull that off. To add to all of this, by the time Deneb releases, i7 will be out, and be even faster still.