How many people want to buy Kentsfield b/c Intel demoed Alanwake?

[dexvx]

Originally posted by: cmdrdredd

Originally posted by: swtethan
i want the Q6600


thats what ill be waiting for


The fact that i wont need a $300 ageia physx card for physics calculations is a pretty good reason.

That's dependant on the game developer. See Alan wake was coded to offload the physics onto a seperate core. If someone wants to support the Physix cards they can code for that.

It's not immediately going to mean that every game from now on will load physics onto a seperate core.

Do you seriously think that the market for Aegiea will be bigger than Quad-Core?

At least with Quad-Core, you can do other stuff (you know general purpose stuff) with that other core. With an Aegiea card, you're stuck with a RISC like interface doing a very specific task. Factor in the cost ($249 for Aegiea, with no sight of a price break) versus the $500-700ish Core2 Quad or even the $999 Core2 Quad Xtreme, and its a non-contest for most people:

E6600 + Aegiea or Q6600

 

[swtethan]

Originally posted by: dexvx

Originally posted by: cmdrdredd

Originally posted by: swtethan
i want the Q6600


thats what ill be waiting for


The fact that i wont need a $300 ageia physx card for physics calculations is a pretty good reason.

That's dependant on the game developer. See Alan wake was coded to offload the physics onto a seperate core. If someone wants to support the Physix cards they can code for that.

It's not immediately going to mean that every game from now on will load physics onto a seperate core.

Do you seriously think that the market for Aegiea will be bigger than Quad-Core?

At least with Quad-Core, you can do other stuff (you know general purpose stuff) with that other core. With an Aegiea card, you're stuck with a RISC like interface doing a very specific task. Factor in the cost ($249 for Aegiea, with no sight of a price break) versus the $500-700ish Core2 Quad or even the $999 Core2 Quad Xtreme, and its a non-contest for most people:

E6600 + Aegiea or Q6600

This is why more game developers will have quad core support for physics, to sell more games.

 

[Pabster]

Originally posted by: swtethan
This is why more game developers will have quad core support for physics, to sell more games.

Of course. But the consumer gets a great benefit too. The awesome physics and special effects are going to be unlike anything we're playing today, and wouldn't be possible without all that horsepower.

 

[harpoon84]

No, one game does not make me want to get a $1000 Kentsfield chip...

It does look impressive though, but I wonder if a C2D chip could achieve similar results. Even if 1 core is totally used for physics, you still have an entire core for the rest of the game.

 

[Cheex]

Core 2 Quad seems very impressive in my opinion. I would rather have 4 cores with one doing physics work rather than an add-in card to do just that.
Makes my computer temperatures higher, takes up space and messes up my airflow...lol. Not to mention the cost.
Well I just recently bought a Pentium D 915 so....when Kentsfield takes the world by storm and Core 2 Duo prices drop...then I'll get Conroe....IF i can't afford Kentsfield...lol...

 

[gobucks]

i'll pass. eventually a quad core will be nice, but by the time it is useful, there will be better quad cores out there. Kentsfield uses the Pentium D approach and just cobbles 2 conroes together. they still have to communicate over an already crowded 1066MHz FSB and the L2 cache is not shared between all 4 cores. I just can't see Kentsfield giving me a noticeable performance boost over my 3.26GHz Core 2 system, especially not $500 worth of extra performance.

I am interested in the next gen parts though - K8L is actually looking a bit nicer than I initially thought; apparently there will be several IPC and branch prediction improvements, and I really like the idea of fully independent core multipliers, which allows you to keep 1 core running at full speed while throttling down the others to reduce power usage. Intel's 2nd gen quadcore (is it nahalem?) will actually be a true quadcore design, will support SSE4, and I'd imagine it will support independence CPU multipliers as well.

 

[dexvx]

Originally posted by: gobucks
i'll pass. eventually a quad core will be nice, but by the time it is useful, there will be better quad cores out there. Kentsfield uses the Pentium D approach and just cobbles 2 conroes together. they still have to communicate over an already crowded 1066MHz FSB and the L2 cache is not shared between all 4 cores. I just can't see Kentsfield giving me a noticeable performance boost over my 3.26GHz Core 2 system, especially not $500 worth of extra performance.

*yawn*

Cobbled FSB? Please. Kentsfield benchmarks over at XS show that they have approximately the same scaling power as 2x DC Opterons in a variety of applications. Noticeable performance boost? Try apps that are optimized for multi-core CPU's (right now its just video editting/encoding and some 3d programs), they give huge performance boosts over core2 Duos. The fact of the matter is, the performance differential is there; its whether or not software dev's choose to use it.

With next generation games, we will see more multi-core action. XBox 360, PS3 Cell are all extremely multi-threaded given their architecture. Its the way gaming is headed. Now, it also depends on other factors. Some start-ups are making RISC based add-on cards (ala Aegiea) to compensate for the lack of cores. I seriously do not see the viability of that given the fact Intel has targetted a $600ish Quad-Core 2.4Ghz in Q1 2007.

 

[cmdrdredd]

Originally posted by: dexvx

Originally posted by: cmdrdredd

Originally posted by: swtethan
i want the Q6600


thats what ill be waiting for


The fact that i wont need a $300 ageia physx card for physics calculations is a pretty good reason.

That's dependant on the game developer. See Alan wake was coded to offload the physics onto a seperate core. If someone wants to support the Physix cards they can code for that.

It's not immediately going to mean that every game from now on will load physics onto a seperate core.

Do you seriously think that the market for Aegiea will be bigger than Quad-Core?

At least with Quad-Core, you can do other stuff (you know general purpose stuff) with that other core. With an Aegiea card, you're stuck with a RISC like interface doing a very specific task. Factor in the cost ($249 for Aegiea, with no sight of a price break) versus the $500-700ish Core2 Quad or even the $999 Core2 Quad Xtreme, and its a non-contest for most people:

E6600 + Aegiea or Q6600

That's not what I said at all stop trying to manipulate what people say and just read what they friggen say. Too many people on here read into what someone said and draw conclusions that have no meaning because it simply wasn't stated the way you think you read it.

Games like UT2007 are using Aegiea. Just because one company, paid by intel, brings a game that does physics and other things on a seperate core does NOT and I stress this... IT DOES NOT mean that every developer will do it. If anything most devs will say it's a waste of time and they'll work on using the GPU more. This is especially true when ATI and Nvidia are trying to get some type of Physics processor on their cards. More people will use those cards than a quad core CPU.

 

[cmdrdredd]

Originally posted by: dexvx

Originally posted by: gobucks
i'll pass. eventually a quad core will be nice, but by the time it is useful, there will be better quad cores out there. Kentsfield uses the Pentium D approach and just cobbles 2 conroes together. they still have to communicate over an already crowded 1066MHz FSB and the L2 cache is not shared between all 4 cores. I just can't see Kentsfield giving me a noticeable performance boost over my 3.26GHz Core 2 system, especially not $500 worth of extra performance.

*yawn*

Cobbled FSB? Please. Kentsfield benchmarks over at XS show that they have approximately the same scaling power as 2x DC Opterons in a variety of applications. Noticeable performance boost? Try apps that are optimized for multi-core CPU's (right now its just video editting/encoding and some 3d programs), they give huge performance boosts over core2 Duos. The fact of the matter is, the performance differential is there; its whether or not software dev's choose to use it.

With next generation games, we will see more multi-core action. XBox 360, PS3 Cell are all extremely multi-threaded given their architecture. Its the way gaming is headed. Now, it also depends on other factors. Some start-ups are making RISC based add-on cards (ala Aegiea) to compensate for the lack of cores. I seriously do not see the viability of that given the fact Intel has targetted a $600ish Quad-Core 2.4Ghz in Q1 2007.

You really think Aegia made their card because there was no core to offload the physics on? My god you are brain dead or something.

Also I don't know about you, but $600 is a damn expensive CPU these days. I spent $220 on my E6400.

 

[gobucks]

Originally posted by: dexvx

Originally posted by: gobucks
i'll pass. eventually a quad core will be nice, but by the time it is useful, there will be better quad cores out there. Kentsfield uses the Pentium D approach and just cobbles 2 conroes together. they still have to communicate over an already crowded 1066MHz FSB and the L2 cache is not shared between all 4 cores. I just can't see Kentsfield giving me a noticeable performance boost over my 3.26GHz Core 2 system, especially not $500 worth of extra performance.

*yawn*

Cobbled FSB? Please. Kentsfield benchmarks over at XS show that they have approximately the same scaling power as 2x DC Opterons in a variety of applications. Noticeable performance boost? Try apps that are optimized for multi-core CPU's (right now its just video editting/encoding and some 3d programs), they give huge performance boosts over core2 Duos. The fact of the matter is, the performance differential is there; its whether or not software dev's choose to use it.

With next generation games, we will see more multi-core action. XBox 360, PS3 Cell are all extremely multi-threaded given their architecture. Its the way gaming is headed. Now, it also depends on other factors. Some start-ups are making RISC based add-on cards (ala Aegiea) to compensate for the lack of cores. I seriously do not see the viability of that given the fact Intel has targetted a $600ish Quad-Core 2.4Ghz in Q1 2007.

are you seriously trying to tell me that current Xbox 360 games actually use all 3 cores, or that PS3 games are gonna use 8 cores any time soon? I'm not disputing that EVENTUALLY quad cores will be useful; i know they will. But i don't think they will be very useful any time within the next year, and in a year there will be better, cheaper quad cores out there.

 

[dexvx]

Originally posted by: cmdrdredd
Games like UT2007 are using Aegiea. Just because one company, paid by intel, brings a game that does physics and other things on a seperate core does NOT and I stress this... IT DOES NOT mean that every developer will do it. If anything most devs will say it's a waste of time and they'll work on using the GPU more. This is especially true when ATI and Nvidia are trying to get some type of Physics processor on their cards. More people will use those cards than a quad core CPU.

Developers will use whatever has the largest median. I predict that Quad-Core will have a larger median than Aegiea cards. Whether or not it wins the ATI/Nvidia PPU is another story (SLi/XFire is a LOT more common than Aegiea and makes more sense, money wise).

Originally posted by: cmdrdredd
You really think Aegia made their card because there was no core to offload the physics on? My god you are brain dead or something.

Also I don't know about you, but $600 is a damn expensive CPU these days. I spent $220 on my E6400.

Yes I sure do. Aegiea was developed years ago (and brought to market within the last year) because at the time, single CPU's were dominating a dual-cores were relatively rare and expensive. The ramp up for dual-core is far faster than anyone has expected and Quad-Core is, for all practical purposes, "early".

I'm not going to say $600 isn't a damn expensive CPU. However, $300 for an E6600 + $250 for Aegiea or $300 E6600 + $250 for extra GPU for SLi/XFire/PPU or $600 for Q6600, its almost a no-brainer for most people. Anyone with a clue will go the SLi/XFire or Quad-Core more than the Aegiea.

Originally posted by: gobucks
are you seriously trying to tell me that current Xbox 360 games actually use all 3 cores, or that PS3 games are gonna use 8 cores any time soon? I'm not disputing that EVENTUALLY quad cores will be useful; i know they will. But i don't think they will be very useful any time within the next year, and in a year there will be better, cheaper quad cores out there.

Its a known fact that the vast majority current Xbox360 games are mainly using 1 core. They are mostly ports from other systems or the dev's are still getting used to using multiple threads. However, I said the way things are going, multi-threaded games will be the future.

 

[nyker96]

Now that I've seent his demo it's pretty extraordinary physics. Although the game isn't demoing anything else that I find to be innovative which hasn't been done by single core games like Prey. The physics has clearly been superior to current single core games. However, if it cost 600 or 1000 whatever the debut price is, it's just not main stream enough for me to buy yet unless I must play Alan Dale ASAP. However, if both Intel and AMD intro their quades in mass by 2nd half of 2007, it might start to push prices down by Xmas next year.

 

[cmdrdredd]

Originally posted by: dexvx
Yes I sure do. Aegiea was developed years ago (and brought to market within the last year) because at the time, single CPU's were dominating a dual-cores were relatively rare and expensive. The ramp up for dual-core is far faster than anyone has expected and Quad-Core is, for all practical purposes, "early".

I'm not going to say $600 isn't a damn expensive CPU. However, $300 for an E6600 + $250 for Aegiea or $300 E6600 + $250 for extra GPU for SLi/XFire/PPU or $600 for Q6600, its almost a no-brainer for most people. Anyone with a clue will go the SLi/XFire or Quad-Core more than the Aegiea.

Well, there's not a whole lot of people who support the physics card and it will fall into nothingness soon. I predict that most developers will not do any sort of physics offloading until ATi/NV get a standard going on their hardware (likely they'll have competing standards that have similar featuresets).