Michael S (already5chosen@yahoo.com) on 8/9/10 wrote:
---------------------------
>AM (myname4rwt@jee-male.com) on 8/9/10 wrote:
>---------------------------
>>Michael S (already5chosen@yahoo.com) on 8/6/10 wrote:
>>---------------------------
>>>AM (myname4rwt@jee-male.com) on 8/6/10 wrote:
>>>---------------------------
>>>>none (none@none.com) on 8/5/10 wrote:
>>>>---------------------------
>>>>>I picked the one which claims x300 speedup.
>>>>>
>>>>>http://www.opticsinfobase.org/oe/abstract.cfm?uri=oe-17-22-20178
>>>>>
>>>>>"In comparison,we compiled the CPU implementation, tMCimg,
>>>>>using the “-O3” option with the gcc compiler and double-
>>>>>precision computation on a single core of an Intel 64bit Xeon
>>>>>processor of 1.86GHz."
>>>>>
>>>>>So single core, low frequency, double precision, out-of-the
>>>>>box compilation. Enough said.
>>>>
>>>>They got these results using a card available (new) for $80+ these days http://www.google.com/products?q=9800+GT&scoring=p
>>
>>>
>>>>
>>>>> (8800GT no longer avail, 9800GT uses the same G92 chip).
>>>>
>>>>Apparently, $80 won't buy you even the cheapest new Core 2 Duo, let alone a 1.86 GHz Xeon, which can cost as little as $188 (E5502) or as much as $3157 (L7555) http://www.intc.com/priceList.cfm
>>>
>>>That's stupid argument. Video card can't do anything on its own.
>>
>>But neither can any CPU. Right, you can't run a system w/o a CPU, so to make the
>>argument valid we should adjust for the price of the cheapest CPU. But even then,
>>the $45 delta (cheapest Celeron) is just enough to shift to the cheapest Core 2
>>Duo. Keep in mind that component costs (PCB, RAM, all the supporting circuitry)
>>work against the GPU, and do so to a larger extent when low-cost hw like said 9800 GT is considered.
>>
>
>The point is, your "apparently, $80 won't buy you even the cheapest new Core 2
>Duo, let alone a 1.86 GHz Xeon" is bogus,

It's no more bogus than implying that CPU can do something on its own. And while I agree that the cost of the cheapest CPU should be taken into account, there are situations when this delta doesn't matter, e.g. when embedded video is not an option for whatever reason or gfx card is already there (so base price includes the cost of discrete video card as well).

And for the GPU/CPU-agnostic the question of which of the two to get is simply one of which of the two will cost $1 less for needed perf. And this is when being even 1.1x faster at the same price counts.

because $70 above platform's absolute
>minimum do buy you 2x2.93 GHz + no-nonsense cache, i.e. 3.15 times more raw FLOPs
>than what was used as "CPU" reference. Add to that double precision. Then add almost
>certainly non-used SIMD on CPU side. Then add less than stellar compiler (although
>gcc is, probably better than its reputation). And suddenly we are back into 10x
>gain range which, everybody here agree, is sometimes possible. All that still on
>dirt chip CPU hardware rather than $1000 hexacore.

You seem to suggest that utilization of Intel's SSE (the vector insns) *with respective 2x/4x gain* is something that's achieved by flipping a compiler switch and will work in just about any code. In my own experience (dating back many years, but still), and experience of others I come across from time to time, this is not the case by far.

Besides, and closer to the topic, the claimed real advantage of 2.5x-5x of GPUs over CPUs (by Mark Roulo) was not demonstrated so far even for such a small collection of papers. Nor was it shown by David Kanter that these reports (*at least* these 10) are all misleading, bogus, erroneous or fishy in any other manner.