acasx.blogg.se

As high frequency digital signals travel through the chips they start to look more and more like sine waves.

In reality this is no where near the case. We all like to think of digital clocking as pure highs and lows with nearly infinitely fast transitions. Keeping 12 cores in lockstep over a larger area is more timing critical than keeping just four or six cores synched over a smaller area. It is likely an internal timing issue as well. Slowing the clocks down reduces that thermal density. With so many cores running so closely together if they were running at the same speed as fewer cores they’d generate more heat in too small a volume. It’s a thermal density issue on Intel’s part, not Apple’s. See all of the benchmark results in the full article here. I predict that the dual D700s in the top 2013 Mac Pro models will reach or exceed 4000KSamples/sec.” “It obliterates the iMac’s single GPU, but a 2010 Mac Pro tower with a really strong AMD GPU like the 7970 (running only on internal factory PCIe power feeds) comes close. So we used it in our last graph to highlight the advantage of the ‘late 2013’ Mac Pro’s slowest pair of GPUs (FirePro D300s),” morgan reports. “LuxMark is one of the few OpenCL benchmarks that uses multiple GPUs.

“Since the newest iMac and Mac Pro both have PCIe flash STORAGE, they are close in speed with the exception of large sequential reads (playback) where the Mac Pro excels.” “When it comes to CPU crunching, the ‘late 2013’ iMac Core i7 4-core matches or beats the ‘late 2013’ Mac Pro 4-core - at leasts in the tests we used here,” morgan reports. “Thanks to the Other World Computing’s test lab, we have some preliminary benchmarks on the entry level ‘late 2013’ Mac Pro 3.7GHz Quad-Core,” rob-ART morgan reports for Bare Feats.