GT6 Framerate depends on which console you have.

3rd gen CECHKxx 80gb Fat Ps3 here: Minor hiccup/glitch on race starts when personal bgm is on, noticeable performance decrease in cockpit view, other than that it's pretty stable to play...and its framerate got better, especially at cockpit, when i adjusted its settings to sharpen, although at some expense of noticeable screen tearing.

 

Launch PS3...never had the slightest of problems

 

I really don't see how the very same chips are effecting different consoles differently. Even if you shrink the die, the chips logic is still identical, the software could not run otherwise.

What you are missing here is streaming. I have seen textures disappear at bathurst then reappear (its rare though). The PS3 is not fitting everything into RAM, you are getting textures when you need them off the HDD it seems. I think what you are seeing is overhead related to texture management in connection with HDD speed. And your tests correlate with this. I don't have most of the issues you speak of, yet my console is a 40GB with an upgraded HDD to a 7200RPM 500GB. According to your tests my SKU should be bad, but its not.

 

I have the console you listed as the worst. Dealer animations are silky smooth, I never realised people had issues with them. The only difference between the console I have and the one you tested is my HDD upgrade.

Also you must realise that when you are buying a car the game is not only loading that car into RAM, it is also installing it into the HDD for next time.

 

Launch Fat as well. No problems with frame rate at 720 and only one instance of screen tearing, but a noticeable drop in frame rate in 1080 and increased tearing.

320GB Slim. (25xxB) No problem with frame rates at all. Some screen tearing though.

I play mostly in cockpit mode, other times in first person.

 

The thinking is that thermal issues are throttling the chips, because yes, they can all do the same number of ops per cycle and have the same cycles per second - except that the Cell scales the clock speed according to how much work it has to do. Not much to do? Scale back to save power (which equals heat in this case).
The theory is that the temperature monitoring built into the silicon can do the same scaling: getting too hot? Scale it back to reduce power; reduce heat input, reduce temperature, keep things solid-state. Scaling the clocks like that at full load will introduce a slowdown effect, as it'll take longer to do the same ops now.

The key point is that whilst the chips are logically equivalent, thermally they are far from equivalent, even within the same "model" (since some chips will need a higher voltage to hit 3.2 GHz than others, due to variations in the manufacturing process - and power is proportional to the square of voltage). Add to that the various coolers that Sony has used over the different models, which have caused problems for some owners, and it's clear that there is plenty of scope for the thermal balance of any model to vary, and to vary between models, too.

That implies a different thermal sensitivity on a per-console basis in terms of computational load, hardware (inc. cooler) and environmental factors, not least of which: dust. So we're looking at which models are more susceptible to finding themselves operating out-of-spec., thermally speaking.
We may find that the frequency of "thermal imbalance" can be pinned on different things for different models; e.g. perhaps the smaller dies suffer from rare 1 mm (say) bubbles in the thermal paste applied at the factory, which constitute a much larger relative area than with the larger dies (using the same process), and thus cooling is more adversely affected when those bubbles occur - relatively speaking. Such a thing shouldn't be possible (at least, not on that scale), but you get the idea. Maybe the fattys are more susceptible to dust because of the higher thermal loads (again, just an example to illustrate) etc.


The harddrive does indeed play a role, in a "full system" sense, but why would it affect the dealership rendering? Mine developed a mild stutter in the dealerships that wasn't there when I first got the game, and now it's gone again.

It's obviously a very complicated set of interrelated effects, which is why you can't say that such and such model will (always) do X. Rather, what we can potentially achieve is to say: if you're getting X, try Y. If it's possible to characterise in that way, even.

 

Isn't the individual PS3 unit also a variable here? A well used unit can be slower than a pristine one of the same version number.

Have you tested this with multiple units of same version?

I have an old fat PS3 and a slim 320GB which to me play GT6 exactly the same, apart for one thing, the old fat PS3 seems to have a shorter stutter just as the race starts than the new one.

 

@gambleboyen, my lengthy post above hinges on that likelihood, so I think individual consoles do vary, for many reasons.

 

This is a very good question and one I have been wondering about. No I have only tested a single one per model. I do not have access to multiple console of the same model.

I think this is very plausible. Here is a little story.

I got 2x 2001a ps3 slims 120gb with the purchase of 2 bravia TVs back in 2009 as a sale offer.

I kept one in one room and the other in another room. One of the ps3 slims would regularly lock up and require a system file check I mean like a few lockups per week. It also seemed to lag more than the other one in gt5. The other one, never had a single hickup. It ran perfectly in the 3 years I had it. I also seem to recall it ran games smoother.

We sold one of them as it was no longer needed, unfortunately the "good" one was sold. The bad one that would always lockup is the one I ran the 2001a tests on in the OP post.

So yes even two identical model consoles can vary. They were both treated the same. One just ran like a dog. Even swapping out the hard drive had no effect. It kept crashing and I think it was core hardware related. For example it would freeze when loading levels in uncharted and sometimes during fire-fighters.

It also got a lot hotter during game play that the other. Felt like a difference when you felt the heat at the bottom of the console.

Even 2 brand new ds3 controllers can feel different to me in. So its not suprising there can be a difference somwhere in a complex kit like a ps3. Have you guys seen the thermal paste job sony does? Its hit and miss and all over the place. Its plausible due to manufacturing differences in the chips some require more volatge and if the thermal paste job is less ideal than another one, it may run into processing defects sooner than another ps3.

Anyone that thinks all ps3s run games exactly the same are shallow minded IMO. The official ratings of the ps3 cell and GPU may have remained the same throughout the years, but the architecture has been resigned numerous times. Not to mention the entire motherboard layout, components, power supplies voltages, cooling etc.

 

I looked up different NASCAR-class cars and noticed that, in the gallery mode in the dealerships, the 2013 cars had a much worse framerate than the 2011 and older cars running on a 120 GB Slim PS3. Framerate also seems to be worse with newer cars such as the Deltawing. The same applies to tracks new to GT6 such as Brands Hatch and Bathurst - they have framerates worse than tracks that appeared in GT5.

It's clear that PD designed the new content in GT6 with the intention that they run good only on the latest PS3 consoles while ignoring performance concerns on older models. In other words, they must have tested the game only on the new super-slim-spec PS3 dev kits.

 

This is circumstantial evidence. Every PS3 has the same processing power, even the very first gen did not overheat at all so performance should be equal across all consoles under the same test parameters. This is BS.

 

Got the 80gb fat one, got micro freezes in tpp and cocpit mode. Changed the setup: RGB to limited , and sharpen in the options, now the framerate is ok just got a lot of flickering in return.

 

not sure whether it is or not. either way after reading this thread, i ordered a 3002b PS3 Slim 320GB whilst they are still readily available, brand new here in the uk. i had gotten one of the new super slim models just a few months back but man does the new over-plasticy design suck ass. worst thing about it was the noise compared to the 3002b PS3 Slim 320GB version. it has a very cheap feel to it, and that awful awful top loading system just about tops it off.

so is there a difference between the variants? who knows. but i'm happy that i've found a 3002b PS3 Slim 320GB before it is too late

 

I have a 120 GB slim and I’ve had no frame-rate issues, no lock-ups, in any view or track, wight he only technical klunk being a slight moment of lag upon completing a race (or more rarely a lap), but overall still rare and almost transparent while playing. The descriptions of the first page seem far exaggerated…

 

PS3 Slim 320GB (CECH 2504B) here and for what i read in your first post, I have right issues right for what I should find.
I mean, 2504B it's in a half way between 2001 and 3001 you tested and in fact I rarely suffer of tearing and fps drop even at 720p, but still not as much I should find with a 2001. More often, i got a half second freeze when I fail license/missions/ecc.
Gonna follow your advice to put "sharpen" on video options (untill now, i didn't select any of the 2).

 

PS3 slim 250GB ( sorry, don't know the model, but original HD, not an upgrade. )

hmmm, the game is really smooth for me, but i did take the PS3 colour options off, as someone suggested earlier in the thread...

My TV is a Panasonic Viera LED of some description or another and i play the game through the "True Cinema" setting at 100hz....so it looks silky smooth...would that have anything to do with it, or not at all ?

Could well be a placebo effect and I am choosing to notice a difference, but i have to say, the game performs better after turning those XMB colour options off. Used to have lots of problems on sections of Bathurst, but no all silky smooth anywhere on the track...

 

However I have been playing in 30+ degree heat lately and have seen no performance downgrade. PS3's don't get much hotter than mine has been, plus I have that 7200rpm drive in there. It the heat was the cause I would see it.

BTW is there any evidence that the Cell in the PS3 does reduce clock speed when hot? Oh and the dealership rendering, as I said, involved installing the data first into RAM and also onto the HDD. On first viewing of a car the HDD is directly involved.

 

Right that snake told about the "sharpness" option, can u describe me both? I mean, what they actually do exactely, i tried them time ago but didn't see much difference...

 

Ps3 slim (last model before superslim ) Bad framerate and bad screentearing. I'd rather have a constant 30fps than this mess!

 

You should just count yourself lucky that you don't have issues, I guess. As I said, there are numerous factors at play. Ambient temperature is almost always less critical than fouling (dust), as long as the air local to the heat sink is cooler than the die (circulation required!).

There is no direct evidence that the Cell does scale clock speeds specifically according to heat, but most processors do these days. Intel's Core2 is the same age, and has on-die thermal monitoring - both chips were CMOS, too. The Cell definitely can scale its clocks, though - it's a touted power saving feature that, in part, allowed it to win performance-per-Watt benchmarks in supercomputing applications. It'd be trivial (and a no-brainer, it's practically free; an engineer's dream) to use a simple thermal weighting in that clock scaling, since the chip definitely has on-silicon temperature sensors also (see a previous post of mine).


However, the facts stand: some people get slowdown, others don't.
Since the consoles are logically equivalent, what's left?

The harddrive doesn't affect rendering speeds unless it's nicking memory bandwidth required for that rendering - in-game, the streaming will be designed not to interfere in that regard. That's why frames will always be on time, but textures will often pop in much later than a scene change. The game will render empty space before it lets the harddrive have higher priority, and often did in the launch version of GT5. The system-level management of resources clearly avoids bottlenecking at the IO level during gameplay, as it should do.

Besides, the BRD -> HDD transfer occurs only once (and causes obvious, jerky pauses of the order of 200+ ms, not consistently choppy framerates around 20 - 50 ms, as I've seen) - the same is true of saving the game, occasionally (presumably, the engine isn't as scheduling-optimised for HDD writes as it is reads, i.e. streaming). But why is the rendering consistently slow for some people, even when that transfer has already taken place? The game won't render the car until it's loaded off the harddrive, so it's not that.

 

I have a CECH 2504B and lots of tearing and unstable Framerate.

Im sure this has nothing to do with PS3 Hardware Revisions or anything.

Like somebody already said, all PS3 have the same CPU-GPU-RAM-VRAM Setup, so every PS3 should perform the same.

The differnces between the Hardware Revision lies in the BD-Drive, HDD, LAN and WLan, and probably a different Bios version and different Motherboard, which should not affect the FPS. (well no more than 1%)


The reason we have varying FPS has someting to do with the Tracks, Car Models, Lighting Engine, Number of Cars on Track, other Special Effect like Smoke etc.

For example take the Viper SRT10 '06 and the Viper GTS '13.
Very similar cars, but in the dealership, the Viper GTS '13 has noticably lower FPS than the SRT10 '06.
Same goes for Tracks, Grand Valley smooth, Apricot Hill low FPS.


For more info: http://www.psdevwiki.com/ps3/Motherboard_Revisions


I doubt this has anything to do with Thermal Trottling. I doubt the PS3 has Thermal Trottling.
I had a PS3 Fat Model CECHH, which was overheating. This resulted in the Fan going 100% (very loud), but framerate was stable all the time. (Of course with games that have stable FPS)

 

I have similar CECH250xx, and it's smooth even on 1080p Bathurst and Apricot Hill are fine, often do arcade races at Bathurst with rain at 100%, 100% water, 10 AI cars, and still no tearing at all, and consistent 50-60 fps. Gallery view of Viper and Vision GT AMG also smooth, no hiccup. Try the Pozzi Camaro RS Gallery view, mine is smooth, but a friend who has FAT PS3 has considerable stutter when going around the car.

 

I meant anecdotal evidence, there isn't enough data points to conclude of anything. !0 people could do the same as the OP and it would still be insufficient data.

 

I have the CHEC 5201A (Slim w/ Killzone 3 Bundle). No framerate issues 99.9999% of the time. Random stutters during races (once every 15 races or so) but everything else runs smooth as butter. Now I did swap out the drive to a 1 TB because I did not have space for GT6. TV is Toshiba 55'' LED TV. Unsure if using Digital out is lighting the load for the HDMI.

 

Slim 120G (CECH-2107A) here, obvious stuttering at some occasions like the twisty down hill of Bathrust, around start of Le Mans and Spa ... etc.

According to OP, my console should perform well, but it's not. In some other threads, it's said GT5 pushes PS3 hardware even more, which I'm not sure.

I'm sure my console was doing fine in GT5 in frame rate, but GT6 gives me a headache of frame rate issue from day 1.

Some previous comments about HDD and cache etc., can anyone provide further explanation? Or, anyone with slim console and smooth frame rate has done the upgrade to the OE HDD, or something else?

This pretty much demonstrates what I see on my screen:



Well, I don't care the console types. I just hope the frame rate can be improved at those moments. Any info? Thanks a lot.

 

CECH-2001b
250GB Slim

Running off a 28' Hannspree LCD TV at Full HD (1080p)

Runs very smooth, framerate is quite smooth, still acceptable even when many cars are on screen.
Have had NO micro freezes since the 1.02 update.
Screen tearing apparent but nowhere near enough to be an issue. Sometimes not even noticeable.

What I would say, xSNAKEx, is get your hands on a capture card and record a sample replay footage running off of each of the models that you tested, and then upload them together in one video as a comparison to see how each one runs side by side.

 

Essentially the same as this (though running through a different TV, but at the same 1080p resolution). I've only seen minor screen tearing at very few specific places. Menus are silky-smooth and until this thread I'd never even heard that microfreezes were a thing.

 

Let's not bump this thread anymore. Revision of console has no impact on performance.

 

proof?

 

There are numerous things that can cause these issues, but a fresh version of all revisions have the same performance. The only differences are HDD and BRD performances. A few things that could cause this (this is just a few of the possibilities):

1. A well used console with loads of installs and uninstalls usually has more stutters than a new one. Old drives and certain drive revisions might suffer from sector faults which again gives issues.
2. A BRD with weak or dusty lens might cause stutter.
3. A console that has overheated once or more will certainly develop issues. These could be small lockups, screen tearing, colour shifts and more.
4. Not all TVs handles frame drops equally. Some drop frames, lose connection, could cause screen tearing or black out totally.
5. A HDMI cable running parallel close to power lines might catch enough interference to also cause screen tearing and stuttering.
6. Unstable wifi connection could cause glitches even though you are playing offline modes, but still are connected to get bonuses and such.