It's too much of a specialised market. The only way it keeps 2 companies going is by the cards being expensive. The only way it would be worth anyone's time would be to produce on-board GPUs and get companies like Dell to use them in their PCs
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Do you think we'll see a third player enter the high end gpu market any time soon?
- Thread starter 233
- Start date
Intel might licence Imagination Technologies/PowerVR and start using them built into CPU’s/Motherboards. That seems the most likely option bus still unlikely.
Imagination Technologies/PowerVR will never anytime soon move into the PC themselves, someone else will have to get a licence and do it. Assuming this is from gaming point of view as PowerVR are meant to bring out a PC none gaming real time ray tracing card this year.
Imagination Technologies/PowerVR will never anytime soon move into the PC themselves, someone else will have to get a licence and do it. Assuming this is from gaming point of view as PowerVR are meant to bring out a PC none gaming real time ray tracing card this year.
Soldato
- Joined
- 22 Apr 2011
- Posts
- 3,850
- Location
- Nottingham
No I just follow them very closely due to investments and being one of the few remaining old school PowerVR fanboys.
For the discrete PC add-on card market - no. I don't think we'll see any other major players. Discrete add-in cards should still dominate the mid-high end GPU market for the next 5 years or so.
Beyond that, we're looking at increased integration of functionality - i.e. the combination of CPU and GPU into a single piece of silicon. Intel has an obvious advantage here, and AMD are also well-placed. In the coming years, companies like ARM and PowerVR could perhaps make a push towards the desktop market - in which case we could see some competition for Intel and AMD.
I'm not sure what Nvidia will do "post-fusion". They could attempt to acquire an x86 licence and try to build their own "CPGPUs", or they could shift their focus to GPU-compute add-in cards, high-end gaming cards (while they still exist), and other specialist equipment (where they make a large proportion of their profits currently anyway). The other possibility is that they will be bought by a bigger fish like Intel.
Beyond that, we're looking at increased integration of functionality - i.e. the combination of CPU and GPU into a single piece of silicon. Intel has an obvious advantage here, and AMD are also well-placed. In the coming years, companies like ARM and PowerVR could perhaps make a push towards the desktop market - in which case we could see some competition for Intel and AMD.
I'm not sure what Nvidia will do "post-fusion". They could attempt to acquire an x86 licence and try to build their own "CPGPUs", or they could shift their focus to GPU-compute add-in cards, high-end gaming cards (while they still exist), and other specialist equipment (where they make a large proportion of their profits currently anyway). The other possibility is that they will be bought by a bigger fish like Intel.
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 33,188
Ultimately gaming gpu's won't be a part of an APU because of expense, if 99% of cpu's sold don't need the power to run the latest game at super high detail..... they won't have it in there.
There is some potential for more modular cpu design in that in several processes and a few years of more parallel software for cpu's and more software that can use an on die gpu to leverage more power we MIGHT see say 16 core cpu's with a 4 "core" gpu, 16 core cpu with a 16 "core" gpu for gamers, and a 28 core cpu + 4 "core" gpu for serious cpu work, where you end up with cpu's for the serious high end market, a midrange powerful cpu with a basic igp, can do some basic gaming but is more for basic acceleration and then a gamers option rather than buying a discrete card.
But frankly its a crapload of time and effort for a relatively small market who, in general lose nothing but gain functionality by having a powerful discrete gpu in a pci-e slot. IE moving it off die free's up the tdp/cooling/size/power of a gpu when having it on die or in a SOC limits everything.
Intel could buy Nvidia but it would be cheaper to let them die, buy up the parts and the people and start fresh.
Nvidia seem set to go for a SOC/cpu market anyway, x86 could be not a big selling point before APU's become a serious threat to the midrange market anyway. If windows works well with arm support then Nvidia have Tegra...... itself no where near capable of high end gaming but a base to build from. They afaik wouldn't get a x86 licence, but they might not want or need one to go after their own APU's.
However fighting Intel..... for cpu and gpu sales is not something just about any company in the world(at this stage) would choose to do.
There is some potential for more modular cpu design in that in several processes and a few years of more parallel software for cpu's and more software that can use an on die gpu to leverage more power we MIGHT see say 16 core cpu's with a 4 "core" gpu, 16 core cpu with a 16 "core" gpu for gamers, and a 28 core cpu + 4 "core" gpu for serious cpu work, where you end up with cpu's for the serious high end market, a midrange powerful cpu with a basic igp, can do some basic gaming but is more for basic acceleration and then a gamers option rather than buying a discrete card.
But frankly its a crapload of time and effort for a relatively small market who, in general lose nothing but gain functionality by having a powerful discrete gpu in a pci-e slot. IE moving it off die free's up the tdp/cooling/size/power of a gpu when having it on die or in a SOC limits everything.
Intel could buy Nvidia but it would be cheaper to let them die, buy up the parts and the people and start fresh.
Nvidia seem set to go for a SOC/cpu market anyway, x86 could be not a big selling point before APU's become a serious threat to the midrange market anyway. If windows works well with arm support then Nvidia have Tegra...... itself no where near capable of high end gaming but a base to build from. They afaik wouldn't get a x86 licence, but they might not want or need one to go after their own APU's.
However fighting Intel..... for cpu and gpu sales is not something just about any company in the world(at this stage) would choose to do.
Doc, you gotta tell me how it ends!
http://vr-zone.com/articles/revenge...gpu-pcie-card-coming-later-in-2012/14609.html
I expect the desktop card will not do very well at first only really being used for professional work not games. Everything I have seen says the 2nd gen ray tracing card, the one due this year is not a gaming card. But that was before PowerVR said the card was changed to a GPGPU and PowerVR's GPGPU has DX10/11 support. So who knows? The most likely thing is the GPGPU is underpowered for games and at around mobile phone level graphics or just above but massively beats AMD and Nvidia at ray tracing.
Although I am hoping for a date of 2013/2014 and still hope Apple will surprise us this year a more realistic number is over the next 5 years mobiles will get real time ray tracing and this year GPGPU's will very slowly start to take off. Once GPGPU and ray tracing software starts becoming more common on mobiles and takes off we will see the desktop card do much better as software is ported across and the desktop card could be used to program for the mobile devices.
If the desktop card is as good as PowerVR say then I expect someone like Apple or Sony will use it or a similar design in a console for ray tracing. The timeline and extended licences sort off match for the PS4. Assuming the PS4 is 2014/2015. Now Sony have said nothing about this so I don't want to spread rumours. It's just my feeling there is a good chance Sony PS4 will use a PowerVR chip and depending just when it arrives ray tracing. Although if the PS4 is well before 2015 no ray tracing support, 2014 might just be pushing it.
In short over the years we will start seeing more software or parts of software that normally run on the CPU moved over to the GPU. But not all software, the CPU still has a place. PowerVR will not win the desktop market and become popular for desktop gaming any time soon.
Lastly I expect over the next 5 years and more likely sooner than later years PowerVR will start shipping 1 million chips per day. The ray tracing card will steal a lot of thunder/business from Nvidia quadro cards.
EDIT:
PowerVR could be the death of Nvidia indirectly. The main markets Nvidia are moving into to survive PowerVR have pretty much beaten them. Less and less people are needed Nvidia quadro cards as the professional market are finding the inbuilt GPU in the CPU is enough or getting there over the next few years. For the higher end the PowerVR desktop card will make more sense then quadro assuming the numbers we hear are true. Same for mobile Nvidia has so far been massively behind. I can see over the next 10 years a large amount of todays quadro card users not needing quadro. That and PowerVR have so far beaten Nvidia in the mobile market.