Any computer gaming enthusiast has probably seen his fair share of gaming benchmarks in the past, but how accurate are benchmarks determined by recorded playbacks? ExtremeTech’s Jason Cross built both Intel and AMD-based systems and recorded performance based on actual gameplay of six popular gaming titles, using FRAPS to measure performance. High-end systems from Intel and AMD, in this case the Core 2 Duo and Athlon 64 X2 5000+, delivered superb performance, although Intel’s newest architecture takes the cake across the board.
Well, I believe that gaming is not the only thing one will do, thus the performance between the two will not be obvious for the eye. So, which one to pick? I would choose E6600 because of its thermal dissipiation being almost half if not less than half that of AMD.
I have seen some websites recording the CPU temperature of both AMD reched almost 60 degrees celsius while the Intel one reched 35 at peak.
On papers AMD is 89 Watt vs Intel 65 Watt Maximum.
Heat is an important factor for me when buying the CPU because I use 6 drives in the case + nvidia 6000 series GPU + other heat generating devices and my room becomes like a desert of Nivada when I operate the two Pentium 4 workstations at once for an hour. I might buy the core2 due not for performance but for heat reduction.
By the way I had 4 failing HDDs in the past 2 years due to this heat build up; my heat sinks are the best in the market ThermalTake hyper 48 (50$)and Zalman CNPS9500AT (75$). all memory modules are heat sinked also. The cace is from ThermalTake and has 10 fans (2 turbine fans). Room Temperature at summer average 86 F and winter average 65 F.
Edited 2006-08-02 20:50
I think Intel hit the nail on the head with this one.
That was a very nice benchmark. What they did well:
1) Comparable systems, w/ justification of components. (they chose similar-featured CPUs in the $300-350 range, current motherboards, and everything else was EXACTLY the same)
2) multiple runs of the same benchmarks, WITH data that proves the results are consistent.
3) real-world benchmarks are variable, but they controlled for that rather well.
I was most impressed that they showed results from a WoW benchmark which showed nothing. THAT’S honesty.
Not so nice: they chose to compare two dual core CPU at comparable price, that’s nice, but why not include mono-CPU (in the same price range of course) too?
I beleive that for gaming purpose mono-CPU still have the edge and will keep it until games are programmed to become multi-threaded, which is not easy: if memory serves Quake3 tried to be SMP friendly and failed..
Is SSE4 already being used? Don’t think so…