AMD Zambezi news, info, fans !

[gosh]

Ah...hehe well that makes sense, thanks for pointing that out. I just got pretty upset cuz that seems to be the trend w/ many reviews nowadays.

Yes I know! But it is still a very important marker to sell processors.
Battlefield 3 will run just fine with any quad. But if an 8 core bulldozer clocked to 5GHz is going to run the game much faster than i5 or any other intel CPU on 800x600, AMD is home among gamers and that is very important

[GenTarkin]

Yes I know! But it is still a very important marker to sell processors.
Battlefield 3 will run just fine with any quad. But if an 8 core bulldozer clocked to 5GHz is going to run the game much faster than i5 or any other intel CPU on 800x600, AMD is home among gamers and that is very important

I suppose, but Ive also seen the opposite be true where at higher resolutions a particular platform fell behind, where at lower resolutions it was leaps and bounds faster ..
But yes, it would be awesome if what your sayin ends up being true and as we all know, games keep gettin more and more threaded...so maybe AMD's CMT approach will prevail afterall...like they thought it would! - would be freakin awesome hehe!

[gosh]

I suppose, but Ive also seen the opposite be true where at higher resolutions a particular platform fell behind, where at lower resolutions it was leaps and bounds faster ..
But yes, it would be awesome if what your sayin ends up being true and as we all know, games keep gettin more and more threaded...so maybe AMD's CMT approach will prevail afterall...like they thought it would! - would be freakin awesome hehe!

Bulldozer is going to perform very well on higher resolutions when loads are hard on the CPU and not bottlenecked by the GPU. i5-2500 is optimized for light loads. Loads that are predicatble, single threaded 32 bit applications (games that has one heavy render thread). 2500 will probably win them all over bulldozer.

Complicated games that scales well and/or is using a lot av memory is going to be executed a lot better om bulldozer.
L3 cache on 2500 is 12 way associative and 6MB, it has a much lower hitrate comparde to 8MB 64 way associative on Bulldozer. Battlefield 3 seems to use loads of memory and 2500 will have more problem to scale then.
2500 isn't tuned to perform well on these type of loads.

[hyc]

About this:

http://library.madeinpresse.fr/samples/MPqY2Vg2v45Z-f


If Bulldozer performs best in Battlefield 3 it will probably be the CPU gamers choose
Look at this picture:

http://gamegpu.ru/Action-/-FPS-/-TPS...-test-GPU.html

Eh? Why is the top bar the i7-930 at 114fps, when the 1100T got 115fps ?

[xVeinx]

Sorted by minimum frames.

[Voodoo²]

But the bar should be shorter anyway.

[gosh]

Add max and min to get the bar length

[JF-AMD]

Interesting. I was mountain biking with a friend today who works for EA. He was telling me that Battlefield 3 is one of the products that his team is handling. I told him he needs to buy some bulldozers.

[xBanzai89]

Interesting. I was mountain biking with a friend today who works for EA. He was telling me that Battlefield 3 is one of the products that his team is handling. I told him he needs to buy some bulldozers.

I really hope you are hinting at something good here.

[LightSpeed]

Interesting. I was mountain biking with a friend today who works for EA. He was telling me that Battlefield 3 is one of the products that his team is handling. I told him he needs to buy some bulldozers.

This sounds good. Seriously.

[Chickenfeed]

Bulldozer is going to perform very well on higher resolutions when loads are hard on the CPU and not bottlenecked by the GPU. i5-2500 is optimized for light loads. Loads that are predicatble, single threaded 32 bit applications (games that has one heavy render thread). 2500 will probably win them all over bulldozer.

Complicated games that scales well and/or is using a lot av memory is going to be executed a lot better om bulldozer.
L3 cache on 2500 is 12 way associative and 6MB, it has a much lower hitrate comparde to 8MB 64 way associative on Bulldozer. Battlefield 3 seems to use loads of memory and 2500 will have more problem to scale then.
2500 isn't tuned to perform well on these type of loads.

Sorry sounds like total shenanigans if you ask me... Not here to kick up dirt but you are making these claims based on whats on paper and theoretical potential. Last time I checked game performance is largely dependent on IPC at 4 threads using *current* game engines (ie scaling past 4 threads in current generation engines is next to nothing ). Perhaps Frostbite 2 will be one of the first to change that tune. Surely things will progress as time goes on and newer more scalable engines are deployed. We will just have to believe it when we see it

[gosh]

Sorry sounds like total shenanigans if you ask me... Not here to kick up dirt but you are making these claims based on whats on paper and theoretical potential. Last time I checked game performance is largely dependent on IPC at 4 threads using *current* game engines (ie scaling past 4 threads in current generation engines is next to nothing ). Perhaps Frostbite 2 will be one of the first to change that tune. Surely things will progress as time goes on and newer more scalable engines are deployed. We will just have to believe it when we see it

Yes Frostbite is the first game engine I have seen that seems to do normal scaling (like other applications that will use more cpu power). I think that games have been blocked to do this because DirectX 10 (and previous versions) because they have been singelthreaded. Rendering from more than one thread would lower performance. The only way to get performance was to dedicate one thread for rendering and this thread would decide overall performance.

DirectX 11 doesn't lower performance if the game render from more than one thread. It should increase in performance but I have read that there are driver issues about that. But it seemes that the game can divide the work in smaller chunks without loosing performance, and this is how you scale if you want to take advantage of cores avalible. The system will dispatch jobs on free resources, if all are taken they are queued.

If battlefield will be a success it will show other gamedevelopers that this is doable and there are a lot of gamer that will upgrade hardware and that will increase the market.

[hyc]

Add max and min to get the bar length

Something is still wrong with the graph. The legend says min and *average* FPS. For the 1100T's average to be higher than the i7's, even tho its min rate is lower, means the 1100T's *max* FPS was VERY MUCH higher than the i7's.

[gosh]

Something is still wrong with the graph. The legend says min and *average* FPS. For the 1100T's average to be higher than the i7's, even tho its min rate is lower, means the 1100T's *max* FPS was VERY MUCH higher than the i7's.

sorry I meant min and average, didn't read. and yes, the graph dosen't show the numbers as it shoud do. bars should start from 0, now it looks like the difference is more than it is

[BeepBeep2]

Something is still wrong with the graph. The legend says min and *average* FPS. For the 1100T's average to be higher than the i7's, even tho its min rate is lower, means the 1100T's *max* FPS was VERY MUCH higher than the i7's.

That is not true.

What is the mean of this set of numbers?
Example AMD machine:
71
126
130
125
124

Example intel machine:
80
131
117
119
122

...5th grade math

[hyc]

That is not true.

What is the mean of this set of numbers?
Example AMD machine:
71
126
130
125
124

Example intel machine:
80
131
117
119
122

...5th grade math

Obviously it depends on the number of samples, which is unknown here.

Assuming only two samples, then 115 = (X + 71) / 2 and 114 = (Y + 80) / 2, therefore X = 159 (1100T) and Y = 148 (i7).

Still, even with your example, the AMD delivers the higher framerate for the majority of the time, (which is obviously the point of an average score) so again, it's giving the better gameplay.

[BeepBeep2]

Obviously it depends on the number of samples, which is unknown here.

Assuming only two samples, then 115 = (X + 71) / 2 and 114 = (Y + 80) / 2, therefore X = 159 (1100T) and Y = 148 (i7).

Still, even with your example, the AMD delivers the higher framerate for the majority of the time, (which is obviously the point of an average score) so again, it's giving the better gameplay.

But you said:

For the 1100T's average to be higher than the i7's, even tho its min rate is lower, means the 1100T's *max* FPS was VERY MUCH higher than the i7's.

Which is unknown.

[danielkza]

Sorry sounds like total shenanigans if you ask me... Not here to kick up dirt but you are making these claims based on whats on paper and theoretical potential. Last time I checked game performance is largely dependent on IPC at 4 threads using *current* game engines (ie scaling past 4 threads in current generation engines is next to nothing ). Perhaps Frostbite 2 will be one of the first to change that tune. Surely things will progress as time goes on and newer more scalable engines are deployed. We will just have to believe it when we see it

Performance is never dependent on IPC, IPC itself means absolutely nothing without a clock speed.

[JumpingJack]

That pic tells . Test must be made at level where ppl play at. For example my 1100T is using only four (4) cores in BF3.

Yep... Also the data set above shown is not wise to use to demonstrate core scaling. BF3 appears to stop at 4 cores (as you observed).

http://benchmark3d.com/battlefield-3...rmance-preview

What we see here is simply a gross misinterpretation.

[Mechanical Man]

This slide isn't about performance, it's about scaling on the CPU. As you can see Battlefield scales to at least 6 cores, probably 8

No it does not scale over 4.

I have 1100T and i can easily look at task manager showing usage of ~66% That would be easily calculated to 4 cores at max usage. I have DX11 gpu. I'll do some tests on lower resolutions & settings than 1680*1050.

[muziqaz]

Hm, when I played BF3 on my thuban, I saw all the threads being used, not fully loaded, but used. Usage was jumping in every thread from 33% to 60-80%. That's on 1080p with ultra settings with 6970@stock. So topping out at 4 threads is not really true
By the way, it ran 55-60fps all the time, and I mean all the time.

[muziqaz]

No it does not scale over 4.

I have 1100T and i can easily look at task manager showing usage of ~66% That would be easily calculated to 4 cores at max usage. I have DX11 gpu.

Instead of looking at overall %s, check out what is each core doing. Then you will see that every core is being used. Its the same with Dirt 3 as well, or Deus Ex HR.

[Olivon]

Hm, when I played BF3 on my thuban, I saw all the threads being used, not fully loaded, but used. Usage was jumping in every thread from 33% to 60-80%. That's on 1080p with ultra settings with 6970@stock. So topping out at 4 threads is not really true
By the way, it ran 55-60fps all the time, and I mean all the time.

That's not true ... On 1080P Ultra settings, you got 30-40 fps on 6970 stock

[Mechanical Man]

Hm, when I played BF3 on my thuban, I saw all the threads being used, not fully loaded, but used. Usage was jumping in every thread from 33% to 60-80%. That's on 1080p with ultra settings with 6970@stock. So topping out at 4 threads is not really true
By the way, it ran 55-60fps all the time, and I mean all the time.

No, you saw all CORES used, there is lot more threads in there.

Windows switches thread from core to other core, that is why it does not show 4*100% & 2*5% load on taskmanager when game uses only four cores worth of heavy threads.

ps. THERE IS NO ULTRA IN BF3 BETA. HIGH IS MAX. yes u can select ultra, but it does not change anything compared to high.

[muziqaz]

Olivon, when you have your own system with 3.8ghz thuban with stock 6970 doing such low numbers, then maybe I would believe. I don't really care what some website got on their system. I had fraps recording gameplay while I was playing it was 55-60fps. If it was 31fps min I would have noticed it with jerky gameplay. Too bad I deleted that POS of a beta from my PC :/