What made the shitty bulldozer good enough to reach 8400mhz on ln2?
Why can't other processors come even close?
>>60124692
cherry picked chips
32nm transistors are less fragile
retarded marketing tactics to appeal to idiots who think that frequency means everything, same idea that intel had back in the p4 days
>>60124707
I mean technically if you had a bulldozer running at that fucking frequency it would be far better than any intel cpu approaching that price
>>60124707
How good of a measurement is clock frequency nowadays? I mean a CPU has so many more specific instructions, dedicated circuits etc. nowadays, how do you compare it to older models like the Pentium chips?
>>60124767
It means less, but it's still important. I can get my A10-5800k to run at 4.7, but comparing it to stock frequency 1700x, it's about 75% perf single thread.
>>60124723
there's a reason why bulldozer was such a fucking jokes anon
>>60124767
>>60124789
>modern OC
>benefiting single threaded applications
Ya no.
>>60124692
As far as I know AMD expected the chips would clock to super high frequencies in normal use (~5GHz). They hand laid out the chips (no autoplace / autoroute). The design started years (?4?) before the semiconductor process existed. They only had engineering guesses as to what the transistors would be like when they began. Turns out the guesses were wrong; the fab process did not perform as expected when the designs were ready.
iirc, leakage began to be a much bigger problem than expected at 32/28nm.
tldr; the chips could clock to the stratosphere because they were designed to do so. Nobody expected when the design began that the heat generated to be nearly as high as it was.
>>60124707
>same idea that intel has right now
Fixed.
>>60124692
It was supposed to be released with 5ghz base clocks @ 20nm but the fabs shit the bed.