Worth Reading?
+1 [5 votes]
I was always taught that you can’t judge a book by its cover. I’ve learned in recent years that you can’t always judge a piece of hardware by early benchmarks either. However, for those of you that like to know these things as early as possible the first benchmarks of the new AMD K10 have shown up, and they’re not good.
I say not good because the system used was a dual-socket machine with quad-core processors and it wasn’t able to outpace similar systems currently on the market from Intel. In the case of the SuperPi 1M it returned results in 39.657 seconds, which is almost identical to the results from a K8 system running at 2.0GHz.
While the scores aren’t very good, AMD claims that not only were the chips from an older build, the HT link seems to have been disabled, which would have caused an issue with the results. The new chips are due out on September 10th, so I’d wait for the full benchmarks then before making any final decisions on the chips.
AMD K10 benchmarks appears, not that good though [via nordichardware]
I don’t like to use SuperPi to test system performance as this program is not well written.
Instead, I prefer QuickPi (http://home.istar.ca/~lyster/pi.html).
On my laptop which has a AMD Turion TL-60 CPU, it take 1.3seocnds to calculate PI to 1m digits, and 7.56 seocnds to 4m digits.
Remember, when you testing such program, don’t take the time by counting the watch, because, it takes time to print out 4m digits on the screen. Try to redirect the output to a file, and read the total time in the file.
LOL HT is disabled thats the power of the AMD, and you bench marked an 8 core system with A single thread app,thats just stupid. I saw these single thread benches coming I knew people would do that, BENCH IT ON SOMETHING THAT CAN USE ALL 8 COREs
LOOK AT THIS BENCH MARK OF K10
http://www.theinquirer.net/?article=41970