• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Hell has frozen!!!

Well, if this is indeed true, Intel has successfully produced the huge die on 14mm already :

According to Anandtech’s Ian Cutress, the die size of the new Knight’s Landing chips will be around ~683mm^2 – which as any enthusiast knows is absolutely huge (and that on a 14nm node!).
Click to expand...


 
damn that's 100mm^2 over GPU's!

This is what Intel is capable of from a foundry point of view vs. competitors ........ it really is an unfair advantage but that is business.
 
Last edited:
Yeah, these forums wouldn't let me post a direct link to the story, but I'm pretty damn speechless at how far ahead of AMD & nvidia Intel is when they want to be.

Even if the yields are very low, just the fact they already engineered and produced a 14mm chip this size shows they are way ahead and can price it appropriately.
 
Yes, that was the story I was referring to. Looks like Intel had a very good plan in place when they set out to acquire Altera.
 
yeah at first it looked like the Altera buy was a bust, but looking at this, really puts a whole new light on things.
 
Too bad AMD did not demonstrate the Cuda Code translation and recompilation and then comparing the the two versions Cuda on Nvidia hardware to AMD HIP code/compiled and ran on AMD hardware. With some better hardware, hopefully Greenland and Zen next year this will grow well for them.

If one has large investments in Cuda/Nvidia hardware and people knowing the ins/outs I find it hard to believe you will get much movement to something else unless you show the real results and they are good.

As for Intel, they have the money and the tech, maybe some bribes as well to take over this market. Nvidia probably has much more to fear from Intel then AMD at this time.
 
noko said:
Too bad AMD did not demonstrate the Cuda Code translation and recompilation and then comparing the the two versions Cuda on Nvidia hardware to AMD HIP code/compiled and ran on AMD hardware. With some better hardware, hopefully Greenland and Zen next year this will grow well for them.

If one has large investments in Cuda/Nvidia hardware and people knowing the ins/outs I find it hard to believe you will get much movement to something else unless you show the real results and they are good.

As for Intel, they have the money and the tech, maybe some bribes as well to take over this market. Nvidia probably has much more to fear from Intel then AMD at this time.
Click to expand...

I think a better benchmark would be cuda code vs cuda translated to opencl both on one nvidia card. That would show regressions as a function of translation.
 
noko said:
Too bad AMD did not demonstrate the Cuda Code translation and recompilation and then comparing the the two versions Cuda on Nvidia hardware to AMD HIP code/compiled and ran on AMD hardware. With some better hardware, hopefully Greenland and Zen next year this will grow well for them.

If one has large investments in Cuda/Nvidia hardware and people knowing the ins/outs I find it hard to believe you will get much movement to something else unless you show the real results and they are good.

As for Intel, they have the money and the tech, maybe some bribes as well to take over this market. Nvidia probably has much more to fear from Intel then AMD at this time.
Click to expand...


Well I don't think its ready yet for something like this, better not to show something till its ready.

You are right, nV has more to fear from Intel on this front, the new Phi is an absolute beast and if there is any delay with Pascal, they are going to loss a lot of this market and it will go quickly. Pascal seems to be 1 or 2 quarters away and companies aren't going to wait much longer when they can get hardware that is 2 to 3 times faster than older Kapler to do work on. If Intel is able to get a foot hold they will take over this market, they have the muscle to do it.
 
You must log in or register to reply here.
Back
Top