Perhaps. AFAIK, it's some more x86 bloat that supposedly has instructions that will improve "float-intensive" applications.Originally Posted by Sn0wm@n
So I guess it will improve software rendering performance!
(but really, figuring out where a vertex of your polygon is, is exactly the same as figuring out by how much and in which direction to move your physics object).
GPUs are basically custom-built for this type of calculation. Well, not only this type, but geometry is a big part of it.
Last edited by iddqd; 04-26-2010 at 11:10 AM.
Sigs are obnoxious.
The only people this doesn't affect are those with only one GPU in their rig which I blindly estimate to be approximately 98% of peopleIt's lame yes but please don't make it out to be some sort of travesty that will actually impact lots of people.
The travesty is what NV is doing not that its really having a big impact.The only people this doesn't affect are those with only one GPU in their rig which I blindly estimate to be approximately 98% of people
Agreed, we should keep this in the realm of the philosophical where it belongs.
Its stops being philosophical when it has effected some one in a way that is undesirable to them.
Anyway NV has missed the huge opportunity it has had over the years with PhysX.
It could of been the physics acceleration hardware to complement whatever else you had.
Instead they cut support for PPU, older NV cards & use from being used with anything other than another NV GPU.
What a waste.
2 thing on that 1) intractable objects have set paths, things that shoot off from explosions do not. 2) how would u expect to have a multilayer game when things could fall differently on each client, if u dont have things predone or rigged to fall the same way u would have to render it on the server then send it to each client it would be imposable to implement.
and what about just cause 2, bioshock 2 or uncharted 2 (its squeal fest)
5930k, R5E, samsung 8GBx4 d-die, vega 56, wd gold 8TB, wd 4TB red, 2TB raid1 wd blue 5400
samsung 840 evo 500GB, HP EX 1TB NVME , CM690II, swiftech h220, corsair 750hxi
Sorry if you got the impression that I think this is some sort of catastrophe. At the moment there are zero GPU physics games on the market that I care one bit about playing. I'm just a little disappointed in Nvidia for continuing to not play nice with others and for that pettiness to have a negative impact on what I can do with the Nvidia card I purchased.The only people this doesn't affect are those with only one GPU in their rig which I blindly estimate to be approximately 98% of people
Things like this may have a greater effect in the future, however. Both AMD and Intel are integrating GPUs onto their CPUs (and perhaps Nvidia will come out with a mainstream CPU+GPU combo someday too).
So nVidia spent money buying Ageia, developing Cuda, porting PhysX to run on Cuda capable hardware to when it's all done, offer PhysX for free (as most people here want) to it's direct competitor.
I would like to live in that business world where every company spends cash on something to improve it's own product and then runs directly at it's direct competitor to show it and share. It would be so cool.
Are we there yet?
Thats because it all over the place... but it certainly has physics! Just throw a grenade at chairs, watch them move about... shooting bullets at cones, flipping them, right-siding them, standing them all back up, etc.. at such a massive scale! So the animations of bricks might be pre-rendered animations, the debris isn't!
who wants phsyX completely free, they should charge devs to use the platform as all other engines/apis do.
and i dont see why any1 should have to pay to make NV hardware work when their hardware is present. its not like physX is working on the ati card its just using cuda on the NV card. if other things worked like that it would be illegal or completely unused, IE should an intel NIC not work with an amd cpu or should an lsi raid card not work if u had other sata/sas ports or should steam not work if u have EA downloder or anouther games client.
i wish that NV and intel would have to sell or split from agea and havok but it wont happen. it should have been illegal for hardware companies to buy API companies as its a conflict of interests. as now physX is a marketing tool to sell NV cards were with agea it was to be a sold API and have PPU and full cpu support
Last edited by zanzabar; 04-26-2010 at 01:57 PM.
5930k, R5E, samsung 8GBx4 d-die, vega 56, wd gold 8TB, wd 4TB red, 2TB raid1 wd blue 5400
samsung 840 evo 500GB, HP EX 1TB NVME , CM690II, swiftech h220, corsair 750hxi
Its not offering anything for free.So nVidia spent money buying Ageia, developing Cuda, porting PhysX to run on Cuda capable hardware to when it's all done, offer PhysX for free (as most people here want) to it's direct competitor.
I would like to live in that business world where every company spends cash on something to improve it's own product and then runs directly at it's direct competitor to show it and share. It would be so cool.
You need a NV card to run all that you have mentioned & i far as i know NV cards are not free so yes NV gets paid because people still have to buy a NV card.
And no one is asking & expecting Cuda & PhysX to run on ATI hardware.
It seems that you have not got a clue & think that this hack makes CUDA & PhysX run on ATI cards.
And you talk about worlds when yours is not even in the present one.
Last edited by Final8ty; 04-26-2010 at 01:56 PM.
It's not about running physx on ATI cards. It's about Nvidia disabling physx on nvidia cards. That's something their own customers payed for - it's not ATI's responsibility to also pay nvidia so that they won't disable PhysX for Nvidia card owners just because they also have an ATI card in the system.So nVidia spent money buying Ageia, developing Cuda, porting PhysX to run on Cuda capable hardware to when it's all done, offer PhysX for free (as most people here want) to it's direct competitor.
Oh, you mean like how ATI is helping develop GPU accelerated Bullet physics on opencl so even Nvidia customers get something out of it for free?I would like to live in that business world where every company spends cash on something to improve it's own product and then runs directly at it's direct competitor to show it and share. It would be so cool.
And funny how ATI paid for DX11 to be implemented in DIRT2 but yet did not hider its performance for NV cards & im sure it runs better on the new NV cards than ATI. talk about double standards that some NV owners think that is perfectly OK.
Some of them really do have a strange sense of logic.
Luka_Aveiro hit the nail on the head.
Final8ty, take some Business, Micro/Macro Economics classes and you'll fully understand the cut throat nature of the beast. I'm not being sarcastic either. You'd be amazed just how business is run.
The bottom line-It's Nvidias and they have no obligation what-so-ever to give that technology to anyone, regardless of how people feel about it.
I understand Business & there is good & bad.
But you don't understand that no one is asking NV to give ATI anything Because its not running on ATI hardware in the first place neither is that hack making it run on ATI hardware. the hack makes NV software run on NV hardware. & NV is obligated to let users of NV hardware use the features that they paid for & NV practice is not as common at the consumer computer level as you make out because it always seems to be NV is the head lines with its continued bad activities.
Last edited by Final8ty; 04-26-2010 at 02:42 PM.
Flinch I think you need to open your eyes. No one is talking about taking technology from them. They advertise their cards run PhysX. Then let them run PhysX! Don't do silly things like exclude operation based on device ids. These are little games that people can bypass. And now since it is not widespread, who cares anyways. You want people to care at this stage, and you want people to support you.
Until they get hit with a class action for arbitrarily disabling peoples graphics/PhysX cards at least, which I see as just a matter of time. AMD won't take legal action until/if PhysX really takes off, simply because Nvidia's current actions are slowing PhyX adoption.
Let's get one thing clear....
They are NOT giving that technology to anyone!
You still need an nVidia card to run Physx. You cannot run it on ATI. That isn't the problem.
The issue is NVIDIA BREAKING PHYSX ON THEIR OWN HARDWARE simply because another video adapter by ATI is detected in the system. It isn't the ATI card running Physx, it is the nvidia card!!
I don't get what is so hard to see about this.
I'm not trying to yell. I'm just bolding and emphasizing important info that seems to be getting lost/skipped over/ignored/whatever.
The Cardboard Master Crunch with us, the XS WCG team
Intel Core i7 2600k @ 4.5GHz, 16GB DDR3-1600, Radeon 7950 @ 1000/1250, Win 10 Pro x64
It has been a while since college Micro/Macro Econ () but I still feel it's acceptable to hold a difference in opinion about what it's economically possible for a business to do and what's reasonable for a business to do.
Yeah, it's probably not illegal for nvidia to withhold PhysX acceleration from those that payed them for it. But it doesn't mean that we have to be happy or quiet about it.
I payed for this card and one of it's advertised features is being able to run PhysX. I'm not even asking for something for free from Nvidia, I'd just like to be able to run an advertised feature. But if I can't use this 285 after I upgrade then what use is it to me? Might as well toss it, or give it away. And I want to keep using it, that's the worst part. It just seems strange to me that for doing that some people act like I've violated the sacred business code or something. Thanks but I'll go on expecting certain levels of behavior out of businesses, you can expect whatever you want out of them.
If you use DARK BASIC PROFESSIONAL w/DARK PHYSICS(which uses a AGEIA ppu for the physics),and you wanted to update your physics driver your WHOLE PROJECT would have been sabotage.THE GAME CREATORS and AGEIA advertised the software as compatible w/ ATI hardware.This should be grounds for legal action right here hhhmmmmm.
Ok, so NV is not giving the tech to ATI. if I'm understanding correctly people are upset that a NV card can not be used in conjunction with an ATI card.
Guess the legalities would have to be looked at by someone other than forum lawyers.
No it does not need lawyers as it just needs common sense because if business thought that this really was OK then ATI/AMD & Intel could lock out all NV cards on ATI/AMD & Intel motherboards instantly with exactly the same reasons as NV & then some as NV hardware & software is running on ATI/AMD & Intel motherboards & CPUs that NV did not pay for in there development .
Last edited by Final8ty; 04-26-2010 at 03:01 PM.
This isn't as big a development problem as you think. You see, computers are inherently deterministic. So, as long as you keep a few key variables in sync across all the clients, things that shoot off from explosions will behave in exactly the same way on each client. Even if the debris is projected randomly, all you'd need to do is to synchronize the random seed variable. Or just use any variable that you already synchronized as the seed, it doesn't really have to be its own variable.
Such as game time in milliseconds.
It's completely mundane, really.
The more interesting question is "how do you keep people from cheating?". Synchronizing the position of all physics objects can become very network bandwidth-expensive, if you have many such objects. Sure, determinism will generate them in all the same ways, but what stops someone from moving a physics object whose position you do not synchronize?
You could stand on a physics object and change its position in your memory. If the game synchronizes its position (which is expensive), you will desynchronize and be dropped from the server for your trouble. But if it doesn't, you'll basically be able to fly. At least it will appear that way to other players, they don't know that you're just standing on a chunk of debris that you're moving.
Last edited by iddqd; 04-26-2010 at 03:37 PM.
Sigs are obnoxious.
The ppu calculates the positions,calculations cant be diff from one machine to the next lol,1+1=2[send to gpu]=we both get the results of 2.If your trying to find out how to stop cheating just make dedicated servers run dedicated ppu's,so player 1 and player 2 ppu's send the calculations to the server ppu then the server ppu sends the calculations to the players gpu.
Last edited by Hell Hound; 04-26-2010 at 03:50 PM.
Posting Permissions
Reply With Quote
Bookmarks