• 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.

*Breaking News* Intel Larrabee is canned

Not suprised... they've been struggling for a while to get the number of cores going that they need to be competitive*... and their entire approach and mentality just simply won't succeed regardless of how much money they throw at it.


* Simulated benchmarks supposed to go head to head with the 280GTX resulted in performance 4x lower than expected.
 
Meh, they are coming into a market very differently to AMD/Nvidia are, so their approach simply won't be the same.

When Nvidia started the market was massively massively growing for discrete cards, thats just not the case now. Back then a substantially worse card still sold, and developement wasn't at the same levels of cost or complexity at all so a failure wasn't a huge deal, just move on to the next card like Nvidia/Ati did back then.

Now you really need something completely competitive either on price, performance or both or it simply won't sell.

Intel could release a duff product or do whats been expected for a long time and make products that are largely just tested internally for a few generations as they learn the ropes of making a fully fledged fighting gpu for the top end. Thing is they have the pockets, the people and time to do what they want.

I'm not sure exactly what Rroff thinks is bad performance compared to a 280gtx, 1/4 of the performance of Nvidia's top end card right now, on an unreleased not seen not final driver and completely not optimised for card is actually downright impressive.

Likewise at the design and testing stage its far cheaper and faster to make what you'd consider mid or low end parts and simulate high end speed, for testing its the architecture that matters not the final speed. Have 50 5870's, or 200 5470's to test on, which is cheaper and better off the same waifer. The 5470 has all the same functionality as the 5870, just less of it, but when it comes to optimising, testing and theorising that doesn't mean anything.

I wouldn't be surprised in the least if what Intel's playing around with is not the best they can make in terms of speed, architecturally though should be identical.

If Nvidia/AMD were a non presence in the market right now, they couldn't just make any old card release it and see what happened, they'd also need to go through several generations of cards internally before having something ready, thats how it goes when entering established markets.

I also wouldn't take any notice of their intergrated when talking about what they can or can't do, its this simple, the intergrated needed to output a signal, and it does that fine, they've hit all their design goals, any additional features you get are gravy. Their drivers and features might suck, but thats because there was no need for them not too.

They've already been buying up lots of small programming firms that focused on paralel programming, its likely lots of those people, and probably a few stolen off Nvidia/AMD for better wages, will make up a completely new and completely different driver team to the intergrated guys.
 
The future is proabably going to be using many smaller processor units in parallel. Intel are, as mentioned above, moving into this area and i suspect they are learning and biding their time.

Unless materials change radically over the next 10 years then they ( gfx makers ) are going to have to go back down the old 3DFX route and slap several GPUs on a card and have them working in parallel. Not only to reduce cost of chip fabbing but also to combat heat. Of course materials could change to allow greater speeds with less heat/power so its still on open book.
 
Right now, Intel dont really need a big GPU, but things may well change if/when AMD start tossing out CPU's with built in GPU's that co-exist/crossfire with slotin GPUs. At that point intel will be under extreme pressure to have an upgrade path for its cpu/gpu combos, and that wont come from ATi or Nvida.

So Larrabee wont be coming anytime soon.. but who cares, pretty sure that Intel will keep working on developing a gpu which can perform at some point.
 
Rroff said:
I don't think this is a good thing tho - we need a 3rd player in the market to shake things up.
Click to expand...

There is no room in the GPU market for a 3rd player. The R&D cost in designing chips is huge, and it's increasing all the time. ATI barely make profits on their current market share.
 
Err this is old news!

Intel has been distancing itself from Larabee for at least a year. They've been losing people left right and centre and the management basically turned their back on the unit after the ridiculous performance figures.
 
You must log in or register to reply here.
Back
Top Bottom