Maybe by the time I build my build at Christmas AMD can have some of the problems fixed in a semi update. Alot of the performance issues from what I've read is a Bug in the cache that couldn't be fixed in time for launch but will be fixed in piledriver. There is 2 billion transistors on this chip it should be making every Intel chip cry for mercy. but yet phenom 2 is beating it down.
They could just release a 32nm thuban cpu 8 REAL cores or just alot higher clocks.
Maybe by the time I build my build at Christmas AMD can have some of the problems fixed in a semi update. Alot of the performance issues from what I've read is a Bug in the cache that couldn't be fixed in time for launch but will be fixed in piledriver. There is 2 billion transistors on this chip it should be making every Intel chip cry for mercy. but yet phenom 2 is beating it down.
They could just release a 32nm thuban cpu 8 REAL cores or just alot higher clocks.
It feels like the bug in the cache isn't the only issue though.
That shouldn't cause such a massive drop in performance.
It feels like the bug in the cache isn't the only issue though.
That shouldn't cause such a massive drop in performance.
Windows 7 is not optimized for bulldozer yet and didn't the Linux kernel get some patch regarding some cache issue with bulldozer. On top of the issues with the chip not clocking as high as they planned because of some issues with GF.
This might of put it ahead of phenom 2 at least and would have forced Intel to push pricing down a little bit.
Obviously something went horribly wrong lucky AMD graphics division is competent so they can keep the company going hopefully.
Core fusing at first looked like a good idea but seems to have back fired quite badly. Looking at wikipedia they have support for some new AMD only instruction set, I reckon it might end up being able to take better advantage of the hardware than x86 can but you would need quite a specific OS to use it. Linux ports might be able to take full advantage of the hardware then.
IMHO they really need to start getting together (Intel, AMD, maybe IBM, MS and Apple) to form a new architecture that can emulate x86/64 without being unusable (see: ARM) for legacy apps, while being it's own new thing taking an even greater advantage over the hardware to the point where it is a significant performance increase.
x86 and x64, given their age, are hitting diminishing returns pretty fast.
Intel can't form a monopoly anyway, the government might try to bail out AMD or some other company might jump into the market. Wasn't there an act enacted by the government to prevent monopolies? (Which didn't work too well btw)
Intel can't form a monopoly anyway, the government might try to bail out AMD or some other company might jump into the market. Wasn't there an act enacted by the government to prevent monopolies? (Which didn't work too well btw)
If this were true, ATT, Comcast, Verizon, and the like would not have monopolies over such large areas of the country for ISPs.
Would be nice if they started using a CPU architecture that didn't suck ass. We've been stuck with the poorly designed x86 architecture since the 1980's.
And use what exactly? x86 is the only widely used CISC on the planet. I sure as hell am not voting for ARM to take over. (The fact that the ARM standard doesn't require a manufacturer to include all of the features on the chip is disconcerting as well)
x86 and x64, given their age, are hitting diminishing returns pretty fast.
x86_64 (aka AMD64) is actually relatively young. Just like I said above, if you can somehow pull a CISC out of nowhere to replace x86 I'd be all for it...
AMD man, step up your game, I was kind of looking forward to what you had to offer.
Building a new rig soon, looks like the i5 2500k is going to be the better option.
They could just release a 32nm thuban cpu 8 REAL cores or just alot higher clocks.
It feels like the bug in the cache isn't the only issue though.
That shouldn't cause such a massive drop in performance.
Windows 7 is not optimized for bulldozer yet and didn't the Linux kernel get some patch regarding some cache issue with bulldozer. On top of the issues with the chip not clocking as high as they planned because of some issues with GF.
This might of put it ahead of phenom 2 at least and would have forced Intel to push pricing down a little bit.
Obviously something went horribly wrong lucky AMD graphics division is competent so they can keep the company going hopefully.
AMD releases 20 core CPU! Intel is stuck with only 8 core!
*10 years later*
2 year old 8 core Intel processors perform better than AMD 20 core CPU.
in 10 years if our programs still aren't multithreaded enough to make usage of more then 8 cores then all programmers should be punched in the face.
LOL!! top
IMHO they really need to start getting together (Intel, AMD, maybe IBM, MS and Apple) to form a new architecture that can emulate x86/64 without being unusable (see: ARM) for legacy apps, while being it's own new thing taking an even greater advantage over the hardware to the point where it is a significant performance increase.
x86 and x64, given their age, are hitting diminishing returns pretty fast.
COME AT ME BRO!
If this were true, ATT, Comcast, Verizon, and the like would not have monopolies over such large areas of the country for ISPs.
And use what exactly? x86 is the only widely used CISC on the planet. I sure as hell am not voting for ARM to take over. (The fact that the ARM standard doesn't require a manufacturer to include all of the features on the chip is disconcerting as well)
x86_64 (aka AMD64) is actually relatively young. Just like I said above, if you can somehow pull a CISC out of nowhere to replace x86 I'd be all for it...