(10-23-2012, 04:29 PM)Zee530 Wrote: [ -> ]Laptops vs Desktop - The never ending saga
>Never ending saga
>Posts an image of a battle that ended very quickly with one member killing the other
Not exactly the best image you could have used.....
Alright I just read the Guru3D review of vishera and before I move on to other sites I thought I would share an overview.
Pricing at launch is actually good this time around. At $195 these are priced fairly and exhibit no signs of AMDs "newness tax".
Idle power consumption is tied with sandy bridge and higher than ivy bridge.
Full loud power consumption is MUCH higher than sandy bridge and ivy bridge.
Temperatures are still an issue at stock.
Overclocking still exhibits poor power scaling with insane power consumption/heat.
Clock rates are about 7-10% higher on average.
All of the tests were highly multithreaded, even the games that were picked. I was looking at the FX-8350 for the record.
I skipped synthetic because they're pretty useless anyways.
CPU queen math (very branchy integer test): Inbetween 2500K and 2600K
FryRenderer (highly optimized integer based software 3D rendering pipleine): Tied with 2600K
Cinebench: Inbetween 2600K and 3770K
Tied with the 3770K in video/audio encoding tests.
Mandelbrot (double precision floating point heavy, highly optimized): In between the 1100T and 2500K.
Zlib compression (integer and memory intensive): Tied with the 980X, above the 3770K, although only slightly.
VP8 video compression: Inbetween the 2600K and 3770K.
SHA1 hasing algorithm: Beats everything.
Memory Read: Lower than sandy bridge but higher than previous AMD architecturtes. Tied with Gulftown and bloomfield.
Memory Write: Same as above but a bit lower than bloomfield/gulftown.
Game benchmarks: Tied with phenom II. Still getting killed by sandy/ivy bridge.
The benchmarks show it to be 12-15% faster than the FX 8150 across the board (which is lower than the 20% AMD was promising) except in Zlib where for some reason it is 27% faster than the FX 8150. As you can see the IPC has hardly increased at all, only around 2-5%. However single threaded tests will likely show a higher IPC difference.
From what these benchmarks show me the performance per watt crown is still going to Intel by a large margin but in multithreaded tests vishera is doing very well (usually better) in performance per dollar than sandy or ivy bridge.
Now I'm about to look at some single threaded tests.