Penryn 40% faster on games, video...

[Eastcoasthandle]

Wow, Intel actually provided some facts. AMD boys can argue all they want but it's more then the blanket statement made about Barcelona with no benchmark information, no reference to what programs/games/etc they were referencing, etc just a blanket statement. Now that Intel actually provides some proof of their claim all you do is wait for the product release and compare, doesn't it? Oh wait, AMD never provided which CPU nor provided information of of the specs that suppose to be beat Intel, just a freaking blanket statement. You don't have to be a detective to figure that one out!

I mean AMD can't even put their reputation where their mouth is when making blanket BS statements like that. But I guess it's good enough for the fanboy! Because AMD said so, whaa!


Alright, what can Barcelona do? I am not interested in dialogue, show me the benchmarks!!!!

[death metal]

Almost same reaction when the first C2D benchmark was released a year ago, disbelief and all AMDroids just ate their words, calling Intel a faker . Nothing new, same old, same old whiners.

And if anyone forgets, Penryn is still based on Intel® Core™, and yet it still boast improvements on multimedia encoding which is superb due to SSE4, and as much as 40% on HL2, that's just wow. Cinebench is also showing Penryn's muscle, as Cinebench has "linear" impact with CPU speed (not much about FSB) so with just 13% clock speed difference, the gain is so much more.

[7he ]-[0rr0r]

Yes, lets all get excited 40% improved performance in games...
(at 320x240 using integrated graphics 0x anti aliasing, 0x anisotropic filtering and all sliders in general set to the far left)

[madcho]

The average is 25%.

13.5% clocked higher.

=> only 11% average perf clock gain.


With 2*6Mo and new bus @ 1333 it's too low to be a serious agains't K10

[gOJDO]

since when amd slides suck and intel slides rule ?

Since AMD slides have only marketing BS without any usefull data. For example: "40% faster than Clovertown in various workloads". Only one benchmark is more worth and speaks more than all their marketing BS spreaded by the clowns like Henri Richard. Intel are not even mentioning AMD, K8 or K10. Instead they have published data WITH NUBMERS and HARDWARE & SOFTWARE SYSTEM SPECIFICATIONS.

of I forgot, since c2d drew intel fanboys out of their holes

A typical comment from an AMD die-hard fanboy.

Seems there is a good ~10% increase in IPC, at least according to that slide

Depends of the software, some software will not benefit at all, while other can perform 100% faster.

But I am mostly curious of the scaling potential of the chip - how much more Mhz will metal gates bring

The little bird said that OCers will love Penryn.

[gOJDO]

With 2*6Mo and new bus @ 1333 it's too low to be a serious agains't K10

BS!

Show me any DATA of K10 performance.
K10 vs Penryn on desktop system? I bet $1000 on Penryn. It will wipe the floor with 65nm K10. Do you accept?

[largon]

Yes, lets all get excited 40% improved performance in games...
(at 320x240 using integrated graphics 0x anti aliasing, 0x anisotropic filtering and all sliders in general set to the far left)

You obviously don't have a clue what "being GPU bottlenecked" means.

[Metroid]

The little bird said that OCers will love Penryn.

It will maybe overclock to 5.0 G air, taking 50% of overclock

AMD speaks too much for nothing showed yet.

I'm not an Intel nor an AMD fan but for now Intel got my vote until AMD proves against it.

[Hicks]

For crying out loud guys, think about it for a second.

There not comparing similar clocks for a start. 40% faster at what settings, with a 500Mhz higher clocked chip, at 800x600?

You know as well as i, it will be no-where near that in gaming.

[7he ]-[0rr0r]

You obviously don't have a clue what "being GPU bottlenecked" means.

I hope you mean that i obviously do understand due to the roll eyes. the conditions i gave are the way to eliminate the gpu bottle neck, i think that's abundantly clear... i hope.

I always buy low end cpu and high end gpu for that reason.
I just get annoyed when cpu makers make that kind of claim, amd or intel and thats what illicited my comment.

[selectodude]

I'll add to that; I want to see an AMD slide with clockspeed listed...

Slides suck, but this one slide is more useful than all of amd's slides so far.

http://en.wikipedia.org/wiki/AMD_K8L#Timeline

[madcho]

For crying out loud guys, think about it for a second.

There not comparing similar clocks for a start. 40% faster at what settings, with a 500Mhz higher clocked chip, at 800x600?

You know as well as i, it will be no-where near that in gaming.

i think low res too. But with a 8800GTX it's very ....


i run css with all @ full 1600*1200*AA4x*AF16x with x1900xt 512 with 100fps all the time. Why i shoud buy peryn ?

for 40% in low res ?

[selectodude]

i think low res too. But with a 8800GTX it's very ....


i run css with all @ full 1600*1200*AA4x*AF16x with x1900xt 512 with 100fps all the time. Why i shoud buy peryn ?

for 40% in low res ?

Counter Strike: Source isn't a benchmark anymore. This CPU will be great for games like Oblivion or Crysis which are massively parallel and have advanced AI systems.

FFS my friend gets 65fps with a 6600GT maxed out.

[keiths]

I want to see the Conroe scenario played out again. I don't need them to be forsale to be rather convinced, just lots of benchmark results all over like it was for Conroe. So.. it's nice that Intel divulged specifics rather than marketing vagaries like AMD, but it's not enough. It'd also be nice if AMD were to do the same.

[.Tret]

http://en.wikipedia.org/wiki/AMD_K8L#Timeline

Oh lawd!

Selective Reading: Do you have it?

see

Since AMD slides have only marketing BS without any usefull data. For example: "40% faster than Clovertown in various workloads". Only one benchmark is more worth and speaks more than all their marketing BS spreaded by the clowns like Henri Richard. Intel are not even mentioning AMD, K8 or K10. Instead they have published data WITH NUBMERS and HARDWARE & SOFTWARE SYSTEM SPECIFICATIONS.

[madcho]

Counter Strike: Source isn't a benchmark anymore. This CPU will be great for games like Oblivion or Crysis which are massively parallel and have advanced AI systems.

FFS my friend gets 65fps with a 6600GT maxed out.

And why intel use HL2 loast cost in low res ?

1-intel is very stupid
2-intel don't have anything to show in fear or other big games in normal res 1200 or 1600

[Eastcoasthandle]

And why intel use HL2 loast cost in low res ?

1-intel is very stupid
2-intel don't have anything to show in fear or other big games in normal res 1200 or 1600

I'm sorry what did AMD show regarding it's blanket statements about Barcelona again? Can you link me to Barcelona's Fear benchmark, I must have missed that one...

[madcho]

don't dream so much guy's, peryn is for Q1/2008.

And Barcelona is not a desktop chip but an opteron. AMD showed servers benchmarks on it.

[Shintai]

And why intel use HL2 loast cost in low res ?

1-intel is very stupid
2-intel don't have anything to show in fear or other big games in normal res 1200 or 1600

Did Intel show a GPU benchmark? I must have missed it. Low res for CPU benchmark. High res for GPU benchmark.

[vitaminc]

that 40% faster on games benchmark is based on FPS of HL2...

[madcho]

CPU benchmark : 3Dmark 06 cpu, and all 3D in software mode.
GPU benchmark : games.

That's a GPU bench. And intel choose a nvidia card witch cannot be very interesting when we now the quality of the drivers under vista. ... yes the bench is under vista.

if you really believe intel can run games 40% faster, with a die schrinck, new instructions, and some more cache ...

[terrace215]

The average is 25%.

13.5% clocked higher.

=> only 11% average perf clock gain.

You're making a common mistake. Most applications do not scale 100% with core clock. That is, if you boost the speed of, say, the 2.93GHz C2D to 3.33GHz, you will not see the benchmarks go up the full 13.6%. Depends on what is really bottlenecking each app. So when you attempt to "take out" the improvement due to clock, you can't use the whole 13.6%.

Hard to know what the average is. Might be best to analyze each app's response to clock changes. But your current methodology is flawed.

[madcho]

You're making a common mistake. Most applications do not scale 100% with core clock. That is, if you boost the speed of, say, the 2.93GHz C2D to 3.33GHz, you will not see the benchmarks go up the full 13.6%. Depends on what is really bottlenecking each app. So when you attempt to "take out" the improvement due to clock, you can't use the whole 13.6%.

Hard to know what the average is. Might be best to analyze each app's response to clock changes. But your current methodology is flawed.

They don't scale like that if you don't get a higher fsb. And for only 13.5% more, the cpu don't loose so much. try with a kentfield extreme and oc it @ QX6800 and the new one @ 1333*10. for large oc like 100%, this loss will be true. no doubt about it.

Try the same bench with same os, with same graphik card with same drivers. you must get 13% higher with the 3.33ghz, no doubt about it

[JamesBong420]

either way.... cpus have taken a GIANT leap in the past few years... and we are all ready for the next gen change.... including vid cards.....

[Anonymous]

Posting Youtube links in a discussion, and especially one of brunette guys singing and dancing is prohibited by the common sense rule #236 - Perkam