That's a part of the "there is no GPC" theoryYes, you're right. Even more local than a single GPC
Jawed
NVIDIA GF100 & Friends speculation
- Thread starter Arty
- Start date
Funny notion:
People dislike that PhysX is on in reviews, but I heard no such protest when reviews did stuff like this:
http://www.hardocp.com/article/2009/12/23/dirt_2_gameplay_performance_image_quality/4
Or this:
http://www.guru3d.com/article/msi-r5770-hawk-review/15
PhysX is a feature, like DX11, DX10.1 or DX10...like it or not
People dislike that PhysX is on in reviews, but I heard no such protest when reviews did stuff like this:
http://www.hardocp.com/article/2009/12/23/dirt_2_gameplay_performance_image_quality/4
Or this:
http://www.guru3d.com/article/msi-r5770-hawk-review/15
PhysX is a feature, like DX11, DX10.1 or DX10...like it or not
Of course, I know. The question is how are the setup/raster engines coupled to the SMs? I don't buy this GPC structure right now. That nvidia can deactivate single SMs is a strong hint that the GPC structure is not real and only present in some shiny graphs.
aaronspink
Veteran
US prices?
1 year warranty? That's a lot of faith in the quality of the hardware there...
psychoticdream
Newcomer
You're missing the point. If it was PhysX compared to an optimized CPU implementation it would be one thing, but with NVIDIA controlling the CPU implementation of PhysX it's just a joke... they have yet to prove any effort in making the CPU implementation competitive. For all we know it's just busy waiting for 90% of the timePhysX is a feature, like DX11, DX10.1 or DX10...like it or not
Furthermore there's no reason that PhysX couldn't be written to run on ATI's latest cards as well in a portable API (DC or OpenCL) and so the comparison is now down to just what NVIDIA's software developers feel like doing... not particularly interesting or hard to predict.And FWIW people DO complain all the time about "benchmarks" that don't produce exactly the same output. If it's the exact same output in DX9 vs DX10 though for instance, it would seem to be fair game from a user's perspective. No intelligent user is going to use NVIDIA's poor CPU implementation of PhysX.
Last edited by a moderator:
DeadlyNinja
Veteran
These posted yet ? Haven't found the other tests. But on these games it look so close
http://translate.google.com/translat...sl=zh-CN&tl=en
http://image3.it168.com//2010/3/26/4...182cd64c00.jpg
http://image3.it168.com//2010/3/26/1...50a09bcaac.jpg
http://image3.it168.com//2010/3/26/5...5cfaea564d.jpg
http://image3.it168.com//2010/3/26/6...56202fae1f.jpg
The links don't work.
aaronspink
Veteran
its listed, everything else of their site loads fine, but the review doesn't seem to be loading.
edit: and hexus.net is dead!
So it slots exactly where everyone expected with a decently aggressive price.
I do think it's odd that Hexus chose the DX11 path for DIRT2 for only the highest resolution (Fermi favored). Be nice to see DIRT2 DX11 at 19x12.
Unlike DX10/10.1/11, it looks like a bug, it smells like a bugPhysX is a feature, like DX11, DX10.1 or DX10...like it or not
Yep it would be interesting to see a working physics engine on the flat unigine heaven stairs without objects falling trough the stairs.
DavidGraham
Veteran
Yes, the traffic took it down. Good job HEXUS, jumping the NDA queue and getting pummeled by the click-falls.
Slightly funny, but at a best-case scenario Fermi's probably leading by 15% for DX11 in Dirt2.
There seems to be 2 situations that work out great for Fermi:
Setup limitations and tesselation. Funny these 2 are at the opposing sides of performance spectrum.
Farcry2... I'm suspecting super fast AA through NVAPI. Wondering.
Funny notion:
People dislike that PhysX is on in reviews, but I heard no such protest when reviews did stuff like this:
http://www.hardocp.com/article/2009/12/23/dirt_2_gameplay_performance_image_quality/4
Or this:
http://www.guru3d.com/article/msi-r5770-hawk-review/15
PhysX is a feature, like DX11, DX10.1 or DX10...like it or not
What Kyle does is what Kyle does, he does things differently to everyone else so you have to accept that and take it for what it is.
In addition to this these are features which future generations of Nvidia cards are expected to have and they are features which will be implemented on nearly 100% of future graphics cards. You would have a point if Kyle introduced a GTX 285 and then complained that one of the monitors didn't work.
DX11 is a key point of differentiation when your competitors don't implement that feature yet. Nvidia had a choice of whether to implement 10.1 or not, they chose to not implement it. Whereas PhysX is something which would never be implemented on ATI cards as its proprietary technology.
People know what PhysX is. If they want it they know that ATI doesn't support it and they don't need a review to tell them that.
Charlie's claims of clocks and 5% is toast.
Unlike DX10/10.1/11, it looks like a bug, it smells like a bug
PhysX will have more to say than DX10.1
If.
Citing the famous (and complete!) answer of the Spartans to an ultimatum of a conqueror roughly stating that Sparta should surrender, otherwise if he takes the city, his troops will burn it down, rape the women, and all inhabitants will be sold as slaves.
That's conclusion is a bit interesting - allows NV to say 20% faster with a reference when it's nowhere close to 20% faster unless you get up to 25x16 with AA and AF cranked. At more mainstream resolutions (19x12) it's more like 5-10% faster. I suppose one doesn't go after a 480 with a paltry 19x12 monitor?
Similar threads
- Deleted member 2197
- Architecture and Products
