I'm not sure about current designs where both AMD and Intel are putting much more effort into squeezing every last bit of performance out (and thus have those extra pipeline stages, fast-switch transistors, etc) but the Athlon XPs were able to downclock reliably from over 2ghz to 600mhz and were under 10W at that point, and though not all could, it seemed a significant number could downclock to 300mhz as well were the power draw was between 3W to 4W. Not too bad considering that the processor wasn't designed for low power at all (at least not exclusively).
Haven't heard about any Athlon 64s below 800mhz (maybe 600?), but that's still pretty low. AMD has started to differentiate their processors though, with the ones using fast-switching transistors being only the absolute top end processors, slow switching going to the low voltage and mobile, and then seemingly a mainstream range of processors. The mainstream processors are much cheaper, but at the same clock speeds about 10% to 20% less efficient than the slow switching, and 10% to 20% off the top speeds of the fast switching processors. (and likewise, the slow switching ones can't hit the clock speeds of the mainstream ones)
A 750CL produced on 90nm running in the 700MHz region would dissipate a tiny amount of heat which is why I don't believe heat comes into any decisions made by Nintendo if indeed Broadway is based on 750CL and is running at around 700 MHz!
Some A64s can go down to 600 Mhz I believe (800 MHz is very common). Pentium M can do 200 MHz. Both do run those speeds with their power management enabled. Those are the older cores with lower FSB ratings, like Clawhammer and Banias. And some of the Semprons I'd guess.
200 MHz on Banias is not common. Usually Speedstep will drop it to 600 MHz. But I had a Sony note a couple of years ago and it came with Sony apps to control CPU speed. Minimum was 200 MHz fanless from 1400 MHz stock.
Some A64s can go down to 600 Mhz I believe (800 MHz is very common). Pentium M can do 200 MHz. Both do run those speeds with their power management enabled. Those are the older cores with lower FSB ratings, like Clawhammer and Banias. And some of the Semprons I'd guess.
200 MHz on Banias is not common. Usually Speedstep will drop it to 600 MHz. But I had a Sony note a couple of years ago and it came with Sony apps to control CPU speed. Minimum was 200 MHz fanless from 1400 MHz stock.
how do the above cores compared watt-wise to IBM's latest G3s? basically, some measure like a mW/MHz in the targetted clock range would be a good indicator.
I wouldn't be surprised to see them do better than IBM's G3 in performance per watt, but that's really irrelevant. They're not comparable cost wise, AMD and Intel are much more restrictive of their IP, and Nintendo really wanted backwards compatibility. An Athlon 64 may beat the G3 in overall mhz and performance per watt/clock, but I believe a G5 may as well, and Nintendo didn't go for a low clocked version of that either. Isn't the die of even a 128KB L2 cache sempron still substantially larger than Broadway? Nintendo is likely buying Broadway at prices Intel or AMD would only sell processors at if they had millions overstock and had to blow them out.
yes, i do agree with what you point out as strong restrains for nintendo (price & compatibility) and i was not contemplating the possibility for nintendo to pick a better performance part per se, i was trying to make a wattage-wise parallel, and i just wated it to be as fair as possible -- since we may not see an exact match of clocks between the considered cores, i was looking at a mW/MHz measure.
Power consumption doesn't scale linearly with frequency though, power per performance would likely be the best measure to make.
darkblu said:some measure like a mW/MHz in the targetted clock range would be a good indicator.
ergo
otherwise i would generally agree with you that ultimately you want power per performance, but what would you take for definitive performance? - dhrystones? MIPS?
and even if we assume we had the correct performance measure, didn't we already agree that ninty had de fact no options outside of the ppc family for compatibility reasons - so what does that leave us with to compare performance-wise then - moto's G4 and latest high-end IBM's? or a hypothetical 'what-if' G3 part by IBM?
mariner, i don't really think i get what you're saying there.
'their cpu of choice is of particularly low TDP ergo they did not have serious TDP considerations for their application' ? - does not sound right to me.
The GX has 1 MiB cache so it would be a shrink to 90nm with half the cache.A few pages ago you can seen that from some estimations (if they are right) the chip can be up to 22mm^ and if like theafu say it has only 512Kb then it can be very well a upgraded GX/CL.
The CL actually has 256 KiB of L2 cache and it is only 2-way set associative, the GX's is 4-way set associative. That said the information on the CL is a little bit thin ATM, one of the improvements the FX/GX had over the CX was a wider 256-bit bus between L1 and L2 which allowed single-cycle transfers of a whole cache-line so this might have been in the Gekko too (and thus in the CL) but I'm unsure about it. The CL datasheet doesn't mention it and IBM's HotChips presentation of the Gekko didn't mention it either so we might never know about it.If you costumize a GX to be BC with Gekko (and with less L2) it would probably end like a CL, so both info can be right. (IIRC he CL also does have the kind of cache improvements of the GX, right?)
I was thinking these CPUs can offer the comparable performance with low wattage. I don't believe the IGN specifications at all,too many inregularities.
I mean 50W on a 789MHz 750CL more like 7W? What about $250 pricetag? Thier markup cannot be that large. Another wierd irregularity is that Broadway was finished in may of 2006 and mass produced in september 2006, if it was a higher clocked Gekko why release so late. I believe IBM is not incompetant or unefficient and know how to build thier products. Even if said product was manufactured in 2000.Just doesn't sound right.
I am not saying IGN is lying or IGN sources are "making news", but what do you have thats official?
I mean if Nintendo wants a cheaper machine they can use the 970FX(built in 2004) and save the time and money on rebuilding a 1999 chip for 2006-2007to compete(yes nintendo calls Sony and Microsoft competitors)with new tech.A PowerPC 970FX (@ 1.6GHz 16W) and a small quiet fan is not that expensive.
The company already has a strong grip on certain sectors. It makes the processors for all three of the top gaming consoles: Sony Computer Entertainment's PlayStation 3, Microsoft's Xbox 360 and Nintendo's Wii. These improved chips could help it reach even further.
IBM's 750 CL is a 32-bit chip running at 400MHz to 1GHz, and is intended for high-performance embedded applications, including consumer electronics, storage and imaging. The 970GX supports both 32-bit and 64-bit applications running from 1,2GHz to 2,5GHz, making it appropriate for applications in high-bandwidth data processing or algorithmic computation
http://www.power.org/news/pr/view?it...d5d55a73 6f70
The PowerPC 970GX, a follow-on to the PowerPC 970FX, supports both 32-bit and 64-bit operations. It features the same power capabilities as its predecessor, but incorporates twice the integrated L2 cache at 1MB. The range of frequencies for the 970GX is 1.2 to 2.5GHz, enabling the chip to support high-bandwidth data processing and algorithmic intensive computations, making it suitable for communications, storage, multimedia and graphics based devices.
http://www.eweek.com/article2/0,1895,2023931,00.asp
Oh and hi everyone nice board!
I personally don't believe IGN (like fox news,abc,NYC times ect they can have "suggestions" from thier sponsors.) because without hard proof(IBM specification sheet, a techhead that opened it up or hacked it,nintendo's specification sheet.) I am not going to drink IGN's kool-aid yet. But if it is true they have great reporters and sources.
IGN in some cases in regards to Wii specs has been far more accurate than the rampant rumors that turn out to be false or specs that are from outdated SDKs. Only thing I would slight them for is the 16MB edram they reported, but have never really confirmed, where as other aspects what broadway is based from, clock speeds are on the mark from I've been trying to confirm by going to do different sources with accurate info. The interpetation of what Matt knows is what gets under my skin more than anything.
Huh? Performance comparable to what? And where do the 50W come from now?I was thinking these CPUs can offer the comparable performance with low wattage. I don't believe the IGN specifications at all,too many inregularities.
I mean 50W on a 789MHz 750CL more like 7W? What about $250 pricetag? Thier markup cannot be that large. Another wierd irregularity is that Broadway was finished in may of 2006 and mass produced in september 2006, if it was a higher clocked Gekko why release so late. I believe IBM is not incompetant or unefficient and know how to build thier products. Even if said product was manufactured in 2000.Just doesn't sound right.
970 is a large and complex architecture, it's built for high clocks (long pipeline), high ILP and SMT.Flux said:I mean if Nintendo wants a cheaper machine they can use the 970FX(built in 2004) and save the time and money on rebuilding a 1999 chip for 2006-2007to compete(yes nintendo calls Sony and Microsoft competitors)with new tech.A PowerPC 970FX (@ 1.6GHz 16W) and a small quiet fan is not that expensive.
Where did you get your information? Developers?
If you are in contact with Wii developers, could you ask them (if they are at freedom to say this) what are the differences in the Wii developement kit and the commercial unit sold to us?
Broadway being a PPC 970 design would also run contrary to what IBM has said about Broadway being 20% more power efficient than Gekko. The 750CL, as pointed out in various threads, can achieve this, but we've yet to see a 970 that can.