Compaq Presario CQ60-615DX Processor Upgrade

I recently came by a lightly used Presario CQ60-615DX. I love this little computer, it's just extremely slow. I upgraded the RAM to 4GB and have a T7350 laying around that I'd like to put in it to replace the JUNK Celery 900 that it came with.

But before I tear the whole thing apart I'd like to know:
1) Will it support a dual-core processor?
2) Will it support a 1066MHz FSB processor?

I've scoured the HP website but can't find anything.

 

Looking here:

See:
Intel® Celeron® Processor 900 (1M Cache, 2.20 GHz, 800 MHz FSB) with SPEC Code(s) SLGLQ

See:
Intel® Core?2 Duo Processor P7350 (3M Cache, 2.00 GHz, 1066 MHz FSB) with SPEC Code(s) SLB53, SLGE3


It seems like they share a socket spec.

Whether it works on your specific system will depend on the BIOS and possibly the chipset used.

I say give it a try - not only will your performance skyrocket - the notebook should actually run cooler too (35W vs. 25W TDP).

Good luck.

 

What processor do you have? There is no such thing as a T7350.

Why is it "JUNK"?

Yes and yes.

No, not at all. Here Intel is using TDP as a marketing ploy.

 

Trottel, how can it be used as a marketing ploy?

Most people don't compare TDP numbers - right?

Anyway, with a dual core (yeah, I linked to a P7350 not a T7350, oops!) the cpu will get to idle quicker than a single core which will always be 'on' and that will get you a cooler system by itself.

If the TDP numbers are the most that the cpu's can survive/operate at - then, how can they be used as a ploy?

 

P7350, not T7350. I guessed, didn't know for sure. It's the 2Ghz dual core out of my dead G51VX-RX05.

And I call the Celeron junk because, quite frankly, it is. The clock speed is decent, but it's a stripped down single core with 1MB of cache and 800MHz FSB.

Updated the BIOS to the newest version on HP's website, I guess now all that's left is to tear it apart and try it out.

 

To complete the same task, the dual core will be 'on' less, but it will use more power when 'on'. But when doing small tasks that don't require much horsepower, they will both be doing the exact same amount of work at the same time. So, in no situation will a dual core be able to consume less power than a single core, all else being equal.

TDP is the thermal design power, the maximum amount of heat energy the cooling system has to be able to remove from the system. Since the heat energy being given off by the processor is directly related to the energy the processor it is consuming, people make the conclusion that TDP is related to power consumption. But it is only related to it in the same way that TDP is related to the heat given off by the processor. It is the maximum that the system has to be able to handle, not how much energy the processor consumes or gives off as heat.

A dual core running at full load will consume twice the energy of a single core running at full load, all else being equal.

With a lower TDP, the cooling and energy regulation circuitry doesn't have to be as good and people can say that the lower TDP processors are more energy efficient, even if they really don't have a clue how much energy any of the processors consume. Even though the Celeron 900 could be given a TDP much lower than 35w, it isn't for two reasons. The first is that the Celeron 900 is a budget CPU, the cheapest mobile Core 2 that Intel sells. The second is that lower TDP processors are sold for more money than similar processors with a higher TDP. Remember, TDP is a somewhat arbitrary figure Intel gives to its different processor families.

As an aside, there are a couple of things irrespective of TDP that Intel has done to the Celeron 900 that makes it a bit less power efficient during low or no use times. It does not automatically lower its multiplier or decrease its voltage, although it does downclock using the FSB. Although during idle, it still has the advantage of being only one core.

 

Cache and especially FSB are mostly irrelevant. What you basically have is a 2.2Ghz single core vs a 2Ghz dual. The P7350 is clearly far more powerful, but the Celeron is really not bad at all for web browsing, movies, and office tasks. I have mine overclocked to 2.93Ghz as you can see in my sig. I set it to 50% speed in the power options menu one day so that it was running at only 1.45Ghz. I completely forgot about it until a week later when I noticed I was getting some stuttering in a very demanding flash application, so I know even at 2.2Ghz it is far more than enough for light usage.

 

Trottel, I have to be honest, you lost me with your TDP explanation.

With regards to a dual core consuming the same power as a single core cpu doing the same work, sure, if we're comparing the chips on a Windows 3.1 O/S.

On a modern O/S install, with AV software installed, there will always be a real and measurable benefit to a dual core over a single core system - in performance and power/heat metrics.

The single core cpu is never at rest - while the dual core can/will operate at a much lower 'intensity' for much less time to do the same work - faster - at much less heat output too.

This is the whole premise of dual core tech - and on that it delivers. At least in my experience (with Intel CPU's).

 

tilleroftheearth, think about it this way:

Let's say we have a single core processor and a dual core processor that is identical to the single core processor in every which way except that it has two cores. If we are running a process that doesn't require full power from the cpu some cycles will be used and some will be idle. If the process requires X cycles every minute on the single core processor, it will also require at least X cycles every minute on the dual core processor. Then if we have Y cycles spent idle for the single core processor during that time frame, we then have 2Y+X cycles spent idle on the dual core processor during that same time frame. So if we have the same amount of cycles being used by both processors, but one has more idle cycles, the one with more idle cycles, the dual core, will consume more energy/release more heat in the same time frame getting the same work done.

Of course the advantage of the dual core is that it can get more work done in the same amount of time or get the same amount of work done in less time, but the above paragraph shows that under no circumstances will the dual core use less power in the same time frame or use less power to get the same amount of work done as a single core.

 

Thanks for trying to explain Trottel, but I still don't see it.

First, your example is not reflected in real life with real cpu's on real O/S installations.

I think you may be stuck with the first generation dual cores; the Pentium D series.

See:
Intel Dual Core Performance Preview Part II: A Deeper Look - AnandTech :: Your Source for Hardware Analysis and News


But even back then with a very non-optimized architecture a second core only added less than 15% to the power consumption.

I would be very hard pressed to believe that today's processors (Core 2 Duo's and even more to the point, i3's and i5's) are even less efficient than in 2005 when two single cores were basically glued together.

However, I am still open to further persuasion (with appropriate facts).

 

You said that a dual core uses less power and puts out less heat than a single core. But there is just no way for that to be true under any circumstances if they are running the same software.

If the single core has 1 billion clock cycles per second, the dual core would have 2 billion clock cycles per second. Each clock cycle used to do work uses a high amount of energy, and each clock cycle idle uses a low amount of energy. If we have a program using 1 billion clock cycles per second, power consumption of the single core would be 1 billion high plus 0 low. Power consumption of the dual core would be 1 billion high plus 1 billion low. Thus the dual core will use more power. The single core cannot exceed a power draw of 1 billion high, so a program that tries to use more clock cycles per second will not increase power consumption of the single core, but may increase power consumption of the dual core all the way up to 2 billion high and 0 low. This hold true no matter what the values of high and low are, as long as high is more than low.

If you still don't agree with me, could you please explain to me how it could even be possible that adding a second core adds less than 0% power consumption?

If the single core processor is running at 2% and the dual core is running at 1%, they are still doing the same amount of work in the same amount of time, using the same amount of clock cycles to do that work and are thus consuming the same amount of power doing that work. But the missing 99% of the dual core is idle and still using power, just like the idle 98% of the single core. But the idle 99% of the dual core is over twice as big as the idle 98% of the single core! There is no way that the dual core can do the same work in the same time, or in even less time by using more than 50% of its available clock cycles, and use less power than the single core. It is just not physically possible.

This isn't related to the above, but the premise of multi-core processors is that it has become that multiple weaker cores will have a greater processing power than a single more powerful core using the same die space. Multi-core processing doesn't offer any other advantage whatsoever and is actually less efficient, to varying degrees.

 

Using your examples, I agree with your conclusion.

But, as mentioned before, your examples are not based in today's reality.

Adding a second core is not 'free' in regards to idle power usage - but intelligent power gating makes it (almost) effectively so.

A cpu will always output more heat at the peak of its clock limit than at lower frequencies. A dual core, by running at much lower frequencies (same amount of work shared by two cores) than a single core takes advantage of that property by running cooler.

The case you're stating is only for a single, single threaded application (and like I said above, I can agree with your conclusion).

The case I'm stating is that in a modern O/S with a full complement of apps and A/V software, there is never just a single thread loading the cpu at any one time.

If you also believe this state to be true, then maybe you can see how I come to my conclusions too:

The single core cpu will be loaded/struggling at a higher frequency and for much longer than a dual core system would. This translates into running at a higher inefficiency for a significantly longer period of time (the single core) vs. running at a greater efficiency (the dual core) for a much shorter period of time (where one or both cores are at 'idle' while the single core is still sweating heavily).

I can see what you're trying to present (finally), but knowing that a single core cpu can easily take 4+ times longer to do the same (modern) work as a dual core is something I've seen again and again with many clients systems compared to my dual+ core setups.

As an example, I'm currently running 102 processes right now at a 0%-5% cpu load (i3 350M cpu). With a single core running less than half as many processes (same basic install + A/V, 0.2GHz slower single core cpu) the cpu was spiking up at idle to 40% constantly and never, ever 'idled' less than 8%.

In a very limited example as you've set up, you are right. In the more real world usage I have seen different cpu/platforms/systems running side by side on, I think the facts are on my side?

When you compound the above with the fact that most single core cpu's do not downclock when idle (as you have stated previously), the dual cores are even more impressive from a (lower) power consumption and a (less) heat generated point of view.

Given that the Celeron 900 in the OP's system is such a CPU (no advanced speedstep circuitry to make it idle more efficiently) maybe you can see why I stated that the change to the P7350 would positively affect how cool the notebook would run compared to the 10W higher TDP of the Celeron processor.

Again, I don't see how the TDP is a 'marketing ploy' by Intel.

The benefits are real and tangible. In smaller, more powerful and cooler running notebooks and in longer run times while on battery power.


On the desktop side:

I can remember, when idling, the computers would heat up the office by themselves (great in the winter). Now, if I don't set the furnace to fire up, the office can be a great imitation of a walk in cooler.

(Above: I'm comparing 8 single core computers to a dozen quad cores, btw).

 

Long story short, The P7350 is not compatible with the Compaq CQ60-615DX. I borrowed the T9300 from my sister's computer and it worked, so I know it supports dual core. No idea why the P7350 didn't work, but neither laptop would boot with it installed and I know the processor isn't broken.

 

Ah, you're right. Sorry. It uses the GL40 chipset, which is not very commonly used. That chipset supports only up to 800Mhz FSB. Most Core 2 laptops use the GM45 chipset.

I agree. There isn't a lot of cost in terms of absolute power consumption with more cores. But if you have one core idling, it will probably use not much more than half the power of two cores idling, and not much more than a quarter of four cores idling.

You got me. Yes, a dual core of the same architecture running at half the speed of a single core can use less power at high load levels if it is running at a lower voltage than the single core.

The main problem with your experiences is that I doubt it was an apples to apples comparison. Some old single core processor is going to get destroyed by a new one, regardless of how many cores it has. You really cannot do a side by side comparison between them unless they are the same architecture, running in identical systems with identical software. Ideally it would be simply swapping out the processors on the same system. All of what you said is very apt if what you were comparing was a new processor architecture to an old one, but that isn't what we are talking about here. On the same architecture, both otherwise identical processors are going to do the same amount of work using the same amount of energy to get the same job done and there is no efficiency bonus or synergistic effect by having an extra core. One misunderstanding you have is of a single thread or a hundred threads. It really makes no difference at all to a single core processor. On the other hand it is important for a multicore processor, and more important the more cores you have, because then the load can be spread evenly over the cores. If having too many threads on a single core meant anything, even if the processor was at low usage, then six core nehalem would do the work far more efficiently than a dual core or even quad core.

The Celeron 900 does downclock. Unlike speedstep which lowers the mutliplier and slightly lowers the voltage, the Celeron 900 downclocks by FSB throttling.

But look, none of what we have been arguing about has anything to do with TDP. TDP is the thermal design power, the most energy Intel says the cooling system has to be able to remove. The processors reach max heat output at full load. There is no way that a 2.2Ghz single core processor could ever consume more power at full load than a 2Ghz dual core of the exact same architecture. The only reason Intel gave it a 35w TDP was for marketing purposes. Their low TDP models are sold at a premium, but that doesn't jive well with the Celeron 900 being the cheapest mobile Core 2 processor.

That is apples to oranges, tiller. Apples to oranges.

 

Trottel, okay, lets say we're both right (using our own assumptions).

The only thing I know is that there is a way for a 2.2GHz single core to be less efficient than a 2GHz dual core. It is manufactured on a less efficient process (requiring higher voltage to be stable).

On topic;

I am saddened that the OP couldn't get his cpu upgraded.

Could he keep his sisters cpu and put the P7350 in hers?

Ah! Just saw that it wouldn't boot with the P7350 in either computer.


How much does a T9300 cost?

 

Trottel Your theorethical explanation is somehow convincing but what about people reporting that after upgrading to dual core battery life in every day usage actually increase, sometimes significant?. And aren;t You doing wrong assumption when You talking about identical cpu different only of number cores - one vs two. What's my point: if we have 2Ghz 800Mhz 2Mb cache dual core cpu is it equivalent is 2Ghz 800Mhz 2Mb cache 1 core cpu ? or maby 2Ghz 800Mhz 1Mb cache 1 core cpu beacause whats make cpu its not only cores but for example cache. If we have 1 core consuming theorethicly 10W, second core 10W and cache 5W that give us 25W cpu, but cpu with one core and that same cache will give us 15W cpu and even we assume that single core cpu have half cache and that half cache consume half energy - what I dont think we can assume - it give us 12.5W cpu. But I think if we have fairly compare 1 and 2 core cpu they should have that same amount of cache because amount of cache influence on performance in lots of task and programs. At the end we have 25W dual core vs 15W singiel core. And thats mean that 2 core even on full load will not consume 2 times more energy. Or maby my thinking is wrong ? I'm just asking? I'm curious myself how it is actually.

 

I think a large part, if not all, is that the dual cores replacing the single cores were more advanced. Most of all the mobile single core Core 2 processors were early stepping Meroms (65nm). As steppings progressed they became more efficient. There are even some cases of power consumption and overclockability differences between early and later processors of the same stepping, the later ones being better. The jump to 45nm was probably the biggest change in decreasing power consumption. The last steppings of the 45nm Core 2, the 3MB cache R0 and 6MB cache E0 offered very significant power saving improvements over all the earlier 45nm steppings, let alone 65nm chips. As it stands, the only 45nm single core socket P processor is the R0 stepping Celeron 900. So really depending on how big the jump was, the power savings could be huge based on the advancement of the Core 2 microarchitecture.

Also Intel has disabled multiplier downclocking and voltage dropping on their mobile Celerons, and although they still allowed downclocking via FSB, the voltage dropping ability is not in the Celerons' favor as far as power consumption.

What I am not trying to say is that the Celeron 900 is awesome or that the single core Core 2 and beyond processors Intel has released are great or anything like that. All I was trying to say was that despite Intel's arbitrary TDP rating, it isn't exactly power hungry compared to other similar processors and in fact could have a much lower TDP than most other mobile Core 2 processors, and that there are no advantages inherent in a dual core design versus a single core of the same architecture.

You're right that the shared cache should remain constant. In reality the cache uses little power compared to the rest of the processor. Definitely not a negligible amount, but it is not very much. When I said how much it would use I said "a little more" than half compared to the dual core to make up for this. But either way, at full load, where TDP is concerned, the factors greatly affecting it are microarchitecture (amount of cache is included here, but I don't think it makes too much difference), number of cores, frequency, and voltage. So how on earth does one expect that a P9700 (2.8Ghz dual-core, 6MB L2 cache, 28W TDP) uses less power at full load than a T6400 (2Ghz dual-core, 2MB L2 cache, 35W TDP)? I'll tell you why; because Intel's TDP's are semi-bogus.

 

You're mixing your TDP and power consumption with your last statement, although given that it's the same microarchitecture, most people would probably assume that the relationship between TDP and power consumption would be proportionate. I'm not as certain, but let's gloss over that for the time being.

In your particular example, I think any difference in power consumption would be contributed to by a lower voltage range on the P9700 (1.012-1.175 V vs 1.00-1.25 V), as well as possibly the 25% faster FSB (1066 vs 800). The P9700 also came out half a year later than the T6400; we don't know what sort of refinements may have been made to efficiencies within the CPU within that time frame.

Either way, this is starting to wander rather far afield from the original topic. This thread ( http://forum.notebookreview.com/har...es/334009-gl40-chipset-cpu-compatibility.html) may be helpful with the CPU upgrade; there is mention of a successful upgrade of a Compaq CQ62-219WM being successfully upgraded from a Celeron 900 to a T6670.

 

Awesome find, thanks!

 

Intel says that a cooling system for a P9700 has to be able to remove 28W from the processor and that a cooling system for a T6400 has to be able to remove 35W from the processor. Thus, Intel is saying that the cooling system for a T6400 has to be better than the cooling system for a P9700. Since heat produced by the processor is equal to its power consumption, from that we can reasonably assume that Intel is telling us that the T6400 uses significantly more power at full load than the P9700, which, by the way, is provably false.

Power consumption for a processor is directly proportional to the number of cores and the clock speed, and proportional to the square of the voltage.



So you can see, at the top of their voltage ranges and at full load, the P9700 will consume roughly 24% more power than the T6400, yet the T6400 has a TDP 25% higher than the P9700! IE, TDP in this case is complete bull.

FSB does not affect the processor's power consumption. In any case, would it not be logical that a higher frequency would consume more power, and not less?

Yes we do; none. The T6400 uses the R0 stepping, and the P9700 uses the E0 stepping. The steppings are identical except for the R0 stepping having 3MB of L2 cache built on the die, and the E0 stepping having 6MB of L2 cache built on the die. If you think that maybe the E0 stepping has superior power consumption to the R0 stepping, don't. Intel's white papers say so and both steppings are found in both the P and T series processor lines. Also the T6400 was only released 6 months before the P9700, and they were made concurrently as soon as the P9700 was released.

For these reasons, TDP is rarely useful information for end users. It is part guide to follow for manufacturers, and part marketing hoopla.

 

The bolded above is the point I (theoretically) disagree with. As per your cited article below, the CPU must take the electrical energy consumed and then "dissipate this energy both by the action of the switching devices contained in the CPU (such as transistors or vacuum tubes) and by the energy lost in the form of heat due to the impedance of the electronic circuits." I would argue that the TDP is largely composed of the second half of the quoted statement, and thus if you can improve the impedance/resistance of the electronic circuits, less energy will be "wasted" and you can have a lower TDP, even if you consume the same (or more) electrical power.

You seemed to miss one relevant word in your quoted article, "given". Also you're missing the capacitance of the processor in your equation. If the T6400 and the P9700 were exactly the same, apart from the clock speed and voltage, I would agree with your assessment. Of course, if that was the case, Intel would be much less likely to assign the P9700 a TDP of 28. My argument is that the T6400 and the P9700 do not have the same capacitance, which changes the equation.

This is not to say that Intel's TDPs are exact, you could be completely correct in that Intel has overestimated the actual TDP of a T6400, and they just slapped 35 watts on it because that's "close", but I think the only way for us to be sure would be for actual measurements to be taken.

My point was more that the extra clock speed that the P9700 has seems to be mostly due to the higher FSB; the maximum multiplier of the P9700 is only one-half more than the T6400. I don't overclock, so I really have no idea, but for those people that FSB overclock, does raising the FSB by, say, 25% also raise the power consumed by the same number? Although given other arguments that have been had on how accurate measurements of CPU power consumption are may mean that this will not lead to any useful conclusions.

Actually, the L2 cache is determined by the fact that the T6400 is a Penryn-3M, and the P9700 a Penryn; the Penryn-3M have a smaller die (87 square mm) while Penryn has a larger die (107 square mm). Some Penryn dies did have half the 6 MB cache disabled, to reduce them to 3 MB of cache. This by itself shows that there must be a lot more differences between the 2 processors, given 2 completely different die sizes.

I agree that TDP is rarely useful information for end users, I'm just not sure that as much of it is marketing hoopla as you suggest. Much like, say, the cubic capacity of an internal combustion engine; it has a very definite use, but people will often take it to mean more than it actually does.

 

(bold, by me)...


Trottel,

Yeah, that is where you are going off on a tangent.

Your bolded statement above is not (always) true. Too many other factors come into play that ultimately turns that extra power into heat. Including (as I've mentioned before) better manufacturing (or better optimized) processes.

My direct observations may be apples to oranges for you, but I think I am pretty good at weeding out the significant and the relevant from the chaos called 'real world'.

In a laboratory environment, with very narrow and restricted testing, I agree with most of what you're saying. In actual use those same variables you accuse me of 'comparing apples to oranges with' are coming back to bite you too.


I'll stand by what I said in multiple posts earlier:

if a 2005 era dual cpu accounts for only 15% more power consumption than a single core, then today, I can believe that with all the current tech available (your 'oranges' ), will consume at the most that much 'extra' power.

What we get though in return is performance that is many multiples of the power increase introduced with a second/3/4th core and more importantly (my initial point I made) a significant decrease in the heat output of the system running basic/intermediate tasks - compared to the one core solution.

A major part of this heat reduction is not only the power of the additonal cores which complete a modern O/S's complex (multi-threaded) tasks faster, but also the fact that cpu's are most in-efficient and generate the most heat when they are pushed to their highest clock speed (made worse for single cores when combined with the exponential extra amount of time they spend at 100% to complete the same complex tasks).

You can play with theories and numbers and state MS is just using marketing bs to sell chips to manufacturers, but my view is that the manufacturers that believe that line of thought are also introducing poorly designed/engineered systems that not only risk the longetivity of the components in question, but also impact the performance of the same systems compared to other more properly engineered examples.

HP and Apple are the worst in this regard ime (run way too hot and get throttled way too much to be thought of as 'productivity' hardware at even the low/medium end of the spectrum).

 

I think you're reading into that sentence a little too much. Where do you think all this energy is ending up? It's simple physics. Energy is never created nor destroyed. After those transistors switch, that energy goes somewhere, as heat. I have heard the one argument that power leaves the CPU in the form of electrical signals. That argument leaves a lot to be desired because those signals are a two way street.

I left of that C value because the difference between them would be small and due the fact that I have no idea what those values are. The effect of cache on power consumption is small, so I didn't think it would be significant. But in spite of that, the value for the P9700 could only be higher since the only difference is that it is the same as the T6400, except that it has 4MB more L2 cache than it. So really if there is any error in the calculation, it would be that the P9700's relative power consumption is even higher than the calculations show. Therefore it is a moot point.

The FSB of the T6400 is 200Mhz and the FSB of the P9700 is 266Mhz. The processor's clockspeed is derived as a multiplier of the FSB. They are commonly listed as 800Mhz and 1066Mhz because since the Pentium 4, the data has been sent four times over the FSB each clock cycle, for an effective clockcycle of 800Mhz and 1066MHz, respectively. The value of the FSB is only useful for determining the bandwidth the processor has with the northbridge. The exact multiplier and FSB are inconsequential to power consumption other than the fact that the processor's clockspeed is the product of the two. It doesn't matter if it is 20 x 300Mhz FSB or 30 x 200Mhz FSB.

The R0 stepping I talked about is the Penryn-3M you mention, and the E0 stepping is the Penryn you mention. They are identical besides the extra cache of the E0 stepping processor. The extra die size is entirely due to this extra cache. If you still don't believe me, look at the usage of the R0 and E0 steppings. The P series processors use both R0 and E0 steppings. The T series processors also use both R0 and E0 steppings. There is no correlation between Intel's TDP figures and whether a processor uses the R0 or E0 stepping.

 

I think that was probably my argument from the last time we went through this sort of discussion. And while I agree that it's a 2-way street, the input part of the street is already accounted for in the power draw of the CPU, IMO. From what I understand about transistors, the controlling voltage/current that switches a transistor "on" either drains through a dedicated outlet, or is subsumed in the greater current that is being controlled. Now, assuming we had perfect super-conducting transistors that would be on or off without losing any electricity through resistance (and thus heat), we could make a CPU with a TDP of 0; as all the power and signals that go into the CPU would come right back out, since none would be retained in the system that would need to be dissipated as heat. Of course, we don't have perfect transistors, and thus we have losses that convert into heat. The question we have, of course, is exactly how much goes through, and how much is retained, and I don't know that any of us have the equipment to do the proper measurements... or that those that do are allowed to tell us the results.

C value is not cache, it's capacitance; how much electricity is retained within the circuitry of the CPU. You assume the difference is small, I'm not as willing to make that assumption. You are correct, though, in that we have no idea what those values are, so we have no way of proving one way or another how significant the values are. Also, per my previous argument, if the P9700 is more "efficient" than the T6400, it can consume as much or more power and still run cooler.

{Snip FSB explanation and effect on power consumption}

Ok.

So, 3 MB of cache occupies 25% of the entire die size of the T6400? Wow. The listed TDP of an R0 and an E0 being the same is easier to understand; after all, who's going to complain if a processor runs cooler than it's listed at? I suppose proof would be getting someone with both the R0 and E0 stepping of a T series processor to test temperatures with them and see if they differ. I'm also less convinced that a given stepping means the same arrangement of transistors across processor types. In other words, I don't think an E0 stepping of the P9700 has the same internal arrangement (although it would have the same microarchitecture) as an E0 T9400. What you seem to be saying is that a R0 T6400 has the exact same internal arrangement as an R0 P8700. Actually, for that matter, look at the P8400; there are 5 different versions of it each under the M0 and R0 steppings. This would seem to say that just using the listed stepping isn't going to work in terms of revisions and changes, or maybe just that Intel's sSpec numbers are even more confusing than we thought.