That is exactly the point I was making. So now reality is *slowly* catching up to you.
No the reality is not catching up to me. The onus is on you to prove that the Cell CPU's superiority justified the extra expenditure in the PS3. I just showed you how with a GTX460 GPU, E6400 was 3x slower in a game at Low resolutions than a Core i7. That means unless all games are suddenly 100% GPU limited at 1280x800 or below, the Cell's CPU superiority should have shown up to support smoother minimum framerates in cross-platform games and yet the opposite happens.
From that we can say:
1) Both consoles are mostly GPU limited, thus wasting $230 on Cell chip manufacturing cost into PS3 was not justified;
2) If the Cell's CPU was faster than 360's it didn't manage to consistently provide higher minimum framerates in games (i.e., it was probably slower most of the time since developers didn't bother optimizing the code well enough to take full advantage of the theoretical performance of the Cell);
3) Having a faster GPU is more important than having a faster CPU for consoles. This further supports the view that the Cell was not justified given that it resulted in a much higher power consuming PS3 (
240W power draw) against the 360 (
180W power draw) and added unnecessary costs to the console's price. If the Cell was not used, Sony could have dropped PS3's power consumption significantly, or chosen to allocate the funds more effectively towards a faster GPU since they'd have more capital budgeting room from not wasting $230 on the Cell. Alternatively, Sony wouldn't have been as limited by the power consumption constraints imposed as a result of using a
80W+ Cell processor. Without the Cell, PS3 could have included a much faster GPU, sold at a much more competitive price, resulted in less financial losses to Sony, or freed up power consumption headroom that the Cell ate up, which instead could have been used towards a higher clocked version of the RSX or a faster GPU design overall.
Instead, Sony used the Cell as the primary component in the PS3 with the RSX GPU considered as the secondary component (almost an emergency on their part), which is the exact polar opposite of how a gaming console should be designed for graphics.
You have been linking i7 benches with modern GPUs to show that 720p was a CPU limited resolution on the consoles. That was just moronic. I got sick of listening to you spew your ignorance on the topic so I posted benches from faster GPUs that were still limited at 720p.
You posted benchmarks of GPUs running MAX settings with AA in games. I posted LOW visual settings. My test actually shows how PS3/360 run games @ Low visual settings. In that instance the resolution of 1280x800 is both CPU and GPU limited at the same time. Since the R500 wasn't much faster than RSX, at the very least PS3 should have been able to maintain higher minimums in at least some games because not all console games are 100% GPU limited.
If I was writing satire of an utterly retarded PC Gamer- that exact sentence would likely come out of their mouth.
There you go again, failing to stick to the topic at hand, attacking the poster. You have a lot of class for an Elite Member. You must have a lot of friends in real life with your attitude. :thumbsup:
eDRAm, UMA and a superior GPU are nothing......... there are just different levels of ignorance, it seems on this subject matter you are pushing hard to reach an entirely new level.
Yes, obviously. HD7970M won't have any performance advantage over HD7660GPU. We also now know from you that it's a fact that the Cell >>>> modern GPUs at processing graphics. Obviously, the Cell >>>> modern x86 CPUs (such as Core i7 Sandy Bridge) for running games since you stated that ad nauseum for 6 years. Got it.
Read the book I linked you so you can stop saying things this dumb. Sony helped design the CPU for the 360. If Sony wasn't building Cell, MS wouldn't have gotten the CPU they had in the 360. There has been a book written on the subject, I linked it for you.
What's that have any to do with you proving that the Cell is faster at running games than a modern x86 processor or proving to us that the Cell is more important for graphics than having a faster GPU? Nothing.
I guess that is what it really comes down to. The lead designer for BOTH CPUs said that Cell was better. He stated in the end the systems would end up comparable because of the 360s superior memory layout and faster GPU. But you, who don't know anything at all about processor design nor coding, you know better then he does.
Oh really? So I should listen to the guy who designed the Cell on his honest opinion about how the Cell compares to modern x86 CPUs for running games? No bias there for sure. :sneaky: So far, you still provided 0 proof how a Cell would outperform a Core i7 CPU for games, like literally 0.
I link the top developer in the world, you counter with an accountant.
Not once did I linked an accountant's words. What I linked were quotes from technical developers, artists and so on, people that actually work on designing the games and programming them. BTW, I know you missed this part but Carmack said consoles are horribly underpowered compared to PCs. You may want to reread that 100x. His quote is in this thread already.
Then you claim to link a PC developer claiming that the TriPOWER CPU beat Cell, when what he actually said was a TriPOWER could beat your beloved x86.
He only said the PowerPC could beat x86 in some areas but not overall. Also, it's just in theory and until it happens, it's meaningless. If it will take a programmer 3 years to make a TriCore Power PC outperform a Core i7 that took another programmer 1 hour to do, then it's meaningless.
You can't quote people that know what they are talking about that agree with you, I can.
Is this a serious statement? Did you even read what you just typed?
So I can't provide support for my side of the argument with people knowledgeable on the topic, only you can, or my sources don't count?
The top coders in the world, the guy who actually built *both* processors, even the people you think back your assertion, they agree with me.
Incorrect, every single sources I quoted agree that Core i7 processor is faster than a PowerPC architecture for running games overall, efficiency and per core performance. Even your source Carmack agrees that PS3 and 360 are hopelessly underpowered. Keep living in a dream world that the Cell is 4-5 generations ahead of Core 2 Duo.
Sony's plan was to use Cell in a variety of different applications, then ARM happened. If you hadn't noticed, it took over as the most popular CPU architecture in the world a while ago.
I haven't noticed. I don't use ARM on the desktop or for laptops. What I noticed is that ARM has 0 to do with what you were supposed to show us:
1) How Cell is faster in overall performance vs. a modern x86 processor for games;
2) How Cell is better in performance/watt vs. a modern x86 processor for games;
3) How it's justified that Sony spent $230 on the Cell instead of allocating the budget towards a faster GPU instead to arrive at a more balanced PS3?
4) How the Cell's theoretical Flop performance at all related to real world performance? (case in point Core i3 vs. FX8150 theoretical floating point performance does not translate into game code)
5) How the Cell's complicated design didn't add to high programming costs, inefficiencies and general inability of most game developers to extract that theoretical performance, as result mostly negating whatever advantage the Cell had? (i.e., how is it that game developers took 2-3 years to just get PS3 up-to-speed in terms of graphics with most Xbox360 games, despite PS3 launching 1 year later with supposedly superior CPU)
6) Why Sony's PS3 does not have overall better graphics than Xbox360 6 years later, outside of a handful of games out of 1000? And especially why PS3 has worse performance in almost all cross-platform games, despite the Cell being so much superior not only to x86 processors, but by extension to the Tri-core PowerPC CPU in the 360?
7) Why Sony would be ditching the Cell for PS4 if it was really superior to modern x86 processors in performance/watt, absolute performance and performance/$?
Cell is both less expensive and uses less power for most computational sets against x86.
Already proven wrong on both accounts. The Cell in the PS3 used > 100W of power at load in the fat PS3. The only reason the power consumption dropped over time is due to continuous node shrinks.
Secondly, the Cell architecture has not proven to be better at running most computational code vs. a modern x86 processor, in fact failing completely to capture any interest in the consumer space.
Dev costs are broken down, $7.00 per chip.
That's not what it cost to put the chip into the PS3. Are you this obtuse, or just plain stubborn and naive? Development cost for the Cell means R&D and engineering costs to design the Cell. That's like saying the development cost for Kepler GPU architecture was $3 Billion dollars and then dividing $3B by total # of Kepler GPUs sold to arrive at a per chip cost. That's not what it costs to design and manufacture a GTX680 chip and then sell it in the open market.
The cost of the Cell in each PS3 = Development cost + manufacturing costs (wafer costs, yields, chip testing phase costs, transportation costs, assembly line costs, manual labour costs, warranty costs, etc.). Those costs in total amounted to $230 per each Cell chip in a PS3 per professional financial analyst breakdown. What you provided me is a # you pulled out of thin air with 0 credibility, obviously since it comes from you and not a 3rd party who does it for a living.
Even what you are saying doesn't make sense. You are saying it only cost $7 in total to manufacture and assemble the Cell in 2006 on 90nm node? :sneaky:
It's pretty much impossible to argue any aspect of the Cell with you. My conclusion is:
1) You may have worked closely on some aspects of the Cell;
2) You were directly or indirectly involved and were affected by the Cell in one way or another at your job (for example coding something using the Cell hardware);
3) You have personal attachment to the Cell in some way;
4) You have a strong aversion against everything related to PC gamers and to x86 code / compilers that you find are inefficient and worthless.
The way in which you defend the Cell without actually backing up why it's superior with hard factual data is very odd. You tell me to go buy a buck by a person who designed the Cell and yet countless articles and professional opinions have contradicted the designer himself, confirming the industry wide belief that the Cell didn't live up to the hype, that it's horribly inefficient to code for and that it's ultimately slow.
It also sounds to me like you may have had a career in programming (doubtful though since you don't seem to agree with any of the programmers who have stated that the Cell was terrible to work with) or engineering but you lack complete and total understanding of business strategy, manufacturing costs, development costs, design costs and how the semi-conductor industry actually works. The fact that you proposed that the Cell only cost $7 for Sony to put into the PS3 shows that you really don't understand how companies actually function and what it costs to bring products from the development stage to market.
Really, what you are saying is:
"The Cell is cheaper, more power efficient and faster for games than ANY x86 CPU. Thus, Sony would be idiotic to abandon the Cell in PS4's design for an inferior x86 CPU." That's pretty much it in a nutshell. Of course that also means if Sony abandons the Cell in PS4, everything you have been saying is wrong because no company would abandon a cheaper, more power efficient and faster processor for what is essentially an inferior design, more power consuming and more expensive design (but of course ONLY according to you, not professionals in the industry).
No matter what though, you'll have no argument left when PS4 abandons the Cell.
The minute PS4 launches, your claims of the Cell's (Main Core + multiple unit SPE setup) superiority in performance/watt, performance/$ and absolute performance against a modern PowerPC architecture/x86 CPU will all be proven wrong. Really, the only way anything you are saying can even be remotely justified is if Sony still continues to use the Cell in PS4. I can't wait until PS4/Xbox720 actually launch and neither uses the Cell, so that we never ever have to hear again from you about how awesome the Cell was on paper.
It's pretty amazing to hear from a person who thinks he is more intelligent than anyone else in this thread or our entire CPU forum that Sony had the world's fastest CPU on the planet for computational tasks and games than any x86 processor made since 2006, and yet not a single blue-chip company believed in Sony's Cell. It's amazing that while no company in the world has been able to produce a faster CPU than Intel for games that somehow a company that recently suffered the
biggest loss in its 50-year-history had the world's fastest CPU for computational tasks and videogames all this time!!
You may want think logically for a second about how flawed all of your reasoning is from a business sense if what you say is true:
"David Turek, IBM's Vice-President of Deep Computing, who said that the company would not be developing the next generation of Cell processor."
Facebook
Twitter