Will Conroe make any difference to games?

[Rich]

Some people can and some cant. Ever spoken to one of your non gaming friends who insists that Fear runs fine at 1600x1200 on his GF3? Smooth as butter.... you know the type Its all objective at the end of the day, depends on the individual.

Personally i think its more important to get a constant FPS than a high one. I would rather have 80 all the time than 132 here and there then a drop to 50 in some parts, because you WILL Notice that drop imho.

 

[BubbySoup]

Can you actually tell the difference between 80 fps and 132 fps?

Maybe not, but you can between 15 and 29 which is where Conroe wins out.

 

[Stelly]

You remebered correctly:

The Elder Scrolls IV: Oblivion
1600x1200, Full HDR, 8xAF, Everything Maximum
Opteron 165 @ 3GHz: Min 15 | Avg 83.993 | Max 689
Conroe @ 3GHz: Min 29 | Avg 132.776 | Max 1546

What graphics card is being used??

Stelly

 

[Vegeta]

7900GT's in SLI 700/800.

 

[Stelly]

7900GT's in SLI 700/800.

How do you run SLi on a Crossfire motherboard??

Stelly

 

[Vegeta]

Don't know, this is the motherboard hes got the 7900gt's running on.

Intel D975XBX "Bad Axe" Mobo

Intel Core 2 Duo (Conroe) E6600 ES @ 3ghz / Max Clocks TBD
2x eVGA 7900 GT SLI @ 700/800 - Forceware 84.56
2x 1GB G.Skill DDR2-800 4-4-4-12
Intel D975XBX "Bad Axe" Mobo
Western Digital 250GB 7200RPM SATA HDD
OCZ GameXStream 700w PSU

 

[Stelly]

Don't know, this is the motherboard hes got the 7900gt's running on.

Intel D975XBX "Bad Axe" Mobo

Intel Core 2 Duo (Conroe) E6600 ES @ 3ghz / Max Clocks TBD
2x eVGA 7900 GT SLI @ 700/800 - Forceware 84.56
2x 1GB G.Skill DDR2-800 4-4-4-12
Intel D975XBX "Bad Axe" Mobo
Western Digital 250GB 7200RPM SATA HDD
OCZ GameXStream 700w PSU

Yes I get that... but how??

Stelly

 

[killer_uk]

Hacked drivers

From what I hear a guy over at XS does them.

 

[Vegeta]

Theorhetically it should work, and it does

 

[The Old Man]

I don't get it, why are there comparisons between an Operton (server processor) and the Conroe (a desktop processor) in games benchmarks? Surely the AMD64 is better at games that an Operton?

 

[Kesnel]

s939 Opterons and A64s are the same architecture. They differ only in name.

 

[The Old Man]

Aha, thanks for that.

 

[drunkenmaster]

why do people give credit to those numbers they look like they can't be possibly real, accurate. maybe he made a mistake maybe not, at gpu limited resolutions stuff like this doesn't happen.

as the, was it fear shows, the amd can at lower resolutions produce 140fps or something average, thats pretty much the "average" fps the cpu can ever produce, which means anytime the framerates drop from higher resolutions then its the gpu limiting it. this has shown to be the way for every game, every cpu tested, every gpu tested for the last 5 years that i've been following hardware. so why is it the gpu's can produce 80+ fps according to that basic review, the opteron according to that same review can pump out over 140fps quite easily, yet its only doing 40fps where the conroe system has 80fps. this is the kind of result that firstly, has been the complete and utter opposite to what every other "preview" has shown, and goes against logic. the numbers are almost useless.


for gaming any cpu will show almost no difference. at the gpu's limits, whatever resolution that will be a fx62 and a 2800+ 754 wouldn't show a very big difference at 1Ghz difference(almost). likewise all other previews show that at higher resolutions the same thing occurs, the conroe isn't pulling very far ahead at all(within a few %) over a fx62, which is within a few % of the lower 939 cpu's aswell. gaming, despite what most want to believe is all down to gpu with pretty much any current cpu. hell, if someone bought a new conroe system with a x6800, and then also spent £50 or borrowed a 2Ghz celery and showed the difference most of you would be shocked. fine at 1024x768 the difference would really be apparent, but at gpu's limit there would still be little in it. unless you really use your cpu, encoding, rendering that kind of thing, most of the time a cpu upgrade will offer you very little.

 

[Durzel]

There, there.. I'm sure AMD will bring out something in a year or so that will be competitive. No need to get upset.

 

[easyrider]

I am already seeing big boosts in games with my conroe over my opty.

 

[drunkenmaster]

I am already seeing big boosts in games with my conroe over my opty.

how big a boost, have you managed to up the resolution that you used to play at with the new system, and what exactly was the old system compared to the current one?

There, there.. I'm sure AMD will bring out something in a year or so that will be competitive. No need to get upset.

not sure tbh if you're being sarcastic of not(damn thee internets ability to truly communicate meaning ) , i think my point was, despite being a little slower IPC and obviously overclocking less well, if amd can get both cores working for any single threaded app then any current am2 will for now with most games using one core, kick conroe's ass in those games and anything else thats single threaded. then comes the 4x4 which again, if amd can get single threaded stuff working across cores will spank conroe very badly. then kentsfield in jan-ish will rock but is already showing massive bandwidth limitations. amd's quad core could do very very well(even without anti-hyperthreading) due to bandwidth limitations and overall core architechture added to the newer higher IPC k8L cores. amd's future is not looking half as bad as it did 3 months ago when we were less than optimistic amd could do anything to fight conroe/kentsfield off.

 

[Durzel]

I was being sarcastic.. but it's not as much fun when the other person doesn't get worked up about it

To be honest though you're basing a lot on "ifs" and "buts".. "if" AMD get multi-core emulating as single core, etc. There's no point talking about AM2 or AMD quad cores beating Conroes/Kentsfields until it actually happens, otherwise we might as well talk about any make-believe chip or technology doing the same. It's easy to speculate about other technology "kicking Conroes ass" when it doesn't exist, and doesn't have to be held up to scrutiny by benchmarks, etc. We might as well say that VIA are gonna bring out a chip that will kick Conroes ass, if we're gonna talk pure speculation based only on fanboy huff and puff.

I prefer to live in the here and now, and there have been enough independant benchmarks that consistently show Conroe will be the fastest thing around for at least the rest of this year and most likely well into 2007. If you were sceptical after the initial Intel/Anandtech benchmarks (and maybe rightly so) then you should - hopefully - be a little more objective about it by now, even if you are an "AMD fan" at heart.

I don't believe in brand loyalty as far as CPUs are concerned, there's no point - you might as well just buy the fastest thing you can afford at whatever time you decide to upgrade. There's no point speculating about the future because if/when AMD do get multi-core working as single core emulation, and whatever jiggery-pokery they have up their sleeve, I'm sure Intel will have something comparable to fight it.. and vice versa, repeat ad infinitum.

Having the sort of blind jingoism for a brand that you seem to have from your previous posts on here is pretty harmful from a "get the best performance I can" perspective. But if you wanna cut off your nose to spite your face, that's your perogative. Me, I'm gonna go with Conroe until I hear of something better.

 

[drunkenmaster]

i know the conroe is EASILY faster, its the "better" architechture right now but benchmarks have already shown a pretty damned significant decrease in speed per core when all cores are being worked even with "just" superpi 1mb which in theory should fit in the cache of the core(8mb on the kentsfield ) . this is something that really the quad core amd cpu won't have, or at least not at all to the same degree. thing is with just the superpi 1 mb i think per core speed went from 20 to 25 seconds roughly, a 25% drop in performance of each core based solely on bandwidth limitations is a huge problem.

fbdimms could be one answer, but so far these chips are supposed to be working on the new boards out now(well gigabytes claim to support it) and fbdimms are looking to be a very expensive way to get mem at the moment.

right now, if you could say every new game would work with multi cores then amd's anti-hyperthreading would essentially be useless and the conroe would win everything hands down. but 2 cores that aren't that much slower are going to easily be faster than one conroe core if thats all the game can use. IF anti-hyperthreading comes out then its a short term fix really for amd which could help it outperform conroe in some areas. but give it a year or two then i doubt many/any games will come out not supporting multi cores, same with most apps. at that point we'll be seeing how improved K8L is, and also how efficient the quad core set up is.

 

[Durzel]

but give it a year or two then i doubt many/any games will come out not supporting multi cores, same with most apps.

Apps definitely, games - doubtful.

As a programmer I can say with 100% certainty that writing games to use multiple CPUs/cores is very, very difficult. It's not beyond the realms of possibility (Q3 Arena did it, etc) but the work involved is not trivial, and the rewards are often not as linear as you might expect, mainly because the sort of things games do (sound, graphics, AI, etc) can't practically run concurrently.

EDIT: That's not to say that 3rd party "snap-ins" like Havok wouldn't benefit from it

 

[Vegeta]

I dont know how people say CPU doesnt make a massive difference, try running your CPU low and then try running your CPU high, theres a world of difference.