| 1 |
On Wed, Jul 25, 2012 at 7:33 PM, Nikos Chantziaras <realnc@×××××.com> wrote: |
| 2 |
> On 25/07/12 23:05, Philip Webb wrote: |
| 3 |
>> |
| 4 |
>> Isn't 22 nm going to be faster than 32 nm ? |
| 5 |
>> [...] |
| 6 |
>> |
| 7 |
>> How do you compare cores vs nm ? |
| 8 |
>> How far is cache size important ( 6 vs 8 MB )? |
| 9 |
> |
| 10 |
> |
| 11 |
> You simply ignore all that stuff and look at how fast the CPUs are. Some |
| 12 |
> 45nm CPUs are faster than some 32nm and 22nm ones. How small the |
| 13 |
> manufacturing process is does not say much about performance. At least not |
| 14 |
> directly. |
| 15 |
> |
| 16 |
> IMO, the best recommendations come from Tom's Hardware. They update their |
| 17 |
> recommendation every month or two: |
| 18 |
> |
| 19 |
> http://www.tomshardware.com/reviews/gaming-cpu-review-overclock,3106.html |
| 20 |
> |
| 21 |
> Ignore the "Gaming" in the title. It's really a recommendation of |
| 22 |
> performance vs price. The only effect their focus on gaming has is that |
| 23 |
> they ignore the integrated graphics of Intel CPUs. |
| 24 |
> |
| 25 |
> |
| 26 |
|
| 27 |
Yes, and remember tomshardware doesn't test Gentoo Linux. Look at a |
| 28 |
site like openbenchmarking.org. |
| 29 |
|
| 30 |
http://openbenchmarking.org/s/AMD%20FX%20-8150%20Eight-Core |
| 31 |
http://openbenchmarking.org/s/Intel%20Core%20i5-3570K |
| 32 |
|
| 33 |
Gentoo results are the most relevant, remember. The bulldozer kind of |
| 34 |
sucks outside of Linux, and even Gentoo, and other benchmarks wont |
| 35 |
show the results you would get. Highly threaded things kick ass on |
| 36 |
the bulldozer, so I imagine updating with it would be fastest. |
Archives