-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
AMD Phenoms to surpass 3 GHz (Follow up)
- Thread starter helmutcheese
- Start date
- Joined
- 19 Jun 2007
- Posts
- 10,717
- Location
- InURmama
Hmm, it was like in 2001/02 I heard about it, not sure what to use a keywords in Google,
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 33,188
well theres lots of info from people who claim a lot of complete nonsense about the press event that amd demo'd the 3Ghz at.
while the press guys that WERE there all have the same story.
yes they were in those silly coolermaster cases with 4x120mm fans on the side panel, they were run with side panel open, so the 4 fans pushed air from one side of an open door to the other, not affecting cooling at all. the quad core 3Ghz used a alu finned small standard amd heatsink with small fan so no massive cooling at all whatsoever. they were allowed to play different games, and benchies just not see or report scores. the guy who ran the event said the parts shown WOULD be available THIS YEAR.
afaik only one extreme edition penryn part is available, most likely december, this year. the rest of the line up is due next year now with only 2 quarters before nehamlem is due. so amd should be getting their chips out before intel. which is a terrible move because, if intel can, which i doubt, get a speed high enough to beat a barcalona then it will still be an extreme edition so £600, if theres a range of speeds for desktop amd k10's then amd get a 2-3 month lead on chips and the last month of the year have lots of chips massively cheaper than intels offering.
intels chip is 25% faster, at 20% higher clock speeds with a 1333 vs a 1066 bus, so in reality its got a 3-4% clock for clock bump, which will only be noticeable in a few benchmarks or if you do a lot of encoding, gaming that won't show up at all.
nehalem is, well at the talk 6 months ago, aimed to bring on die mem controller to the xeons, but with the current massive cache Intel go to that added to on die mem will equate to very expensive chips and there was lots of talk of the P4 days when extreme editions were just xeons, and the desktop normal parts had no on die mem controller. not sure if thats try yet, if it is they lose an advantage.
don't forget that the core 2 duo dedicates a LOT of transistor power to prediction tech, cache, better prefetch and basically a heck of a lot of stuff that already lets it have an incredibly low latency, its basically as fast, latency wise, as a ath 64 already, its hugely hugely better than the P4. an ondie mem controller will REPLACE most of that stuff rather than reduce latency massively further so big increases aren't going to be very apparent. the main difference is bandwidth and the HT bus that gives the amd series a big advantage, even more so on a quad core, that will not change in any way with a on die mem controller. a different/new bus, wider and better and faster like the HT< is what is needed.
i have a quad core, any benchie will tell you, i can get say 14seconds on super pi with one version running. it goes up to around the 16-18second mark per instance if you run 4 instances. thats a 20% performance hit for using all 4 cores, not because of a mem controller, but because the bus is massively limited.
so far , i'll admit i need to read up more, there isn't a whole lot of seen that says nehalem will actually be a big step from a penryn. what i can tell you now is, other than sup com, which i can confirm personally is smoother once you get all the units going, with a quad core, 99.9999999999999999% of games now, and 99.9999999999999999%(i'm not counting, assume thats one less 9
) games by middle of next year will still show no more than a 4% difference in frame rate from a 2.6Ghz ath 64, to a X2 at 2.6Ghz, to a conroe at 2.6Ghz to a quad core at 2.6Ghz.
i'll buy a chip cos its fast, so what, just be aware that it won't improve games massively.
the main reason ath 64's were actually better than the p4's for gaming even with both fairly decent high end cpu's, was the incredibly bad latency on the P4. conroe which fixed the latency stopped that. i've had 2.6-3Ghz single, dual and quad cores. i've played all the normal "big AAA titles" most people have, there is simply not a difference with any of the cpu's except in sup commander, even there 90% of the game is the same, its just smoother with crazy numbers.
while the press guys that WERE there all have the same story.
yes they were in those silly coolermaster cases with 4x120mm fans on the side panel, they were run with side panel open, so the 4 fans pushed air from one side of an open door to the other, not affecting cooling at all. the quad core 3Ghz used a alu finned small standard amd heatsink with small fan so no massive cooling at all whatsoever. they were allowed to play different games, and benchies just not see or report scores. the guy who ran the event said the parts shown WOULD be available THIS YEAR.
afaik only one extreme edition penryn part is available, most likely december, this year. the rest of the line up is due next year now with only 2 quarters before nehamlem is due. so amd should be getting their chips out before intel. which is a terrible move because, if intel can, which i doubt, get a speed high enough to beat a barcalona then it will still be an extreme edition so £600, if theres a range of speeds for desktop amd k10's then amd get a 2-3 month lead on chips and the last month of the year have lots of chips massively cheaper than intels offering.
intels chip is 25% faster, at 20% higher clock speeds with a 1333 vs a 1066 bus, so in reality its got a 3-4% clock for clock bump, which will only be noticeable in a few benchmarks or if you do a lot of encoding, gaming that won't show up at all.
nehalem is, well at the talk 6 months ago, aimed to bring on die mem controller to the xeons, but with the current massive cache Intel go to that added to on die mem will equate to very expensive chips and there was lots of talk of the P4 days when extreme editions were just xeons, and the desktop normal parts had no on die mem controller. not sure if thats try yet, if it is they lose an advantage.
don't forget that the core 2 duo dedicates a LOT of transistor power to prediction tech, cache, better prefetch and basically a heck of a lot of stuff that already lets it have an incredibly low latency, its basically as fast, latency wise, as a ath 64 already, its hugely hugely better than the P4. an ondie mem controller will REPLACE most of that stuff rather than reduce latency massively further so big increases aren't going to be very apparent. the main difference is bandwidth and the HT bus that gives the amd series a big advantage, even more so on a quad core, that will not change in any way with a on die mem controller. a different/new bus, wider and better and faster like the HT< is what is needed.
i have a quad core, any benchie will tell you, i can get say 14seconds on super pi with one version running. it goes up to around the 16-18second mark per instance if you run 4 instances. thats a 20% performance hit for using all 4 cores, not because of a mem controller, but because the bus is massively limited.
so far , i'll admit i need to read up more, there isn't a whole lot of seen that says nehalem will actually be a big step from a penryn. what i can tell you now is, other than sup com, which i can confirm personally is smoother once you get all the units going, with a quad core, 99.9999999999999999% of games now, and 99.9999999999999999%(i'm not counting, assume thats one less 9
) games by middle of next year will still show no more than a 4% difference in frame rate from a 2.6Ghz ath 64, to a X2 at 2.6Ghz, to a conroe at 2.6Ghz to a quad core at 2.6Ghz. i'll buy a chip cos its fast, so what, just be aware that it won't improve games massively.
the main reason ath 64's were actually better than the p4's for gaming even with both fairly decent high end cpu's, was the incredibly bad latency on the P4. conroe which fixed the latency stopped that. i've had 2.6-3Ghz single, dual and quad cores. i've played all the normal "big AAA titles" most people have, there is simply not a difference with any of the cpu's except in sup commander, even there 90% of the game is the same, its just smoother with crazy numbers.
Soldato
- Joined
- 24 Jul 2006
- Posts
- 8,876
- Location
- Hoddesdon, London, UK
BoomAM said:Whats the name?
Wouldnt mind reading up on that CPU..
I think this is a the very article i read, or some variant of it a few years ago
http://www.xbitlabs.com/news/cpu/display/news621.html