Dolphin, the GameCube and Wii emulator - Forums

(03-28-2013, 05:38 PM)admin89 Wrote: [ -> ]

Quote:Amd turion II x4 P540 @ 2.4ghz

It's a dual core CPU
http://www.notebookcheck.net/AMD-Turion-II-P540-Notebook-Processor.37885.0.html

ya i know.... but i have no clue how a 4 got typed in there O.o

So Turion X2 is as fast as Llano x4 clock for clock
Turion destroys Llano in single threaded performance and Llano gains 20% performance from background processing (2 cores for Dolphin , 2 cores for...)
Edit : I forgot to include GPU performance . AMD Mobility 5470 is too slow , it does not meet Dolphin minimum requirement -> It bottleneck the CPU -> Turion x2 is faster than Llano x4 in Dolphin after all

(03-17-2013, 07:36 AM)Garteal Wrote: [ -> ]He actually benchmarked it twice. The first result is exactly the same (save the 2nd save) due to a margin of error.

Quote:and it looks like haswell won't be a big upgrade speed-wise but it'll mostly contain improvements to power management or something like that.

What? Where did you get that from?
As far as I know, the expected increase in performance is around 10% due to the fact that it has 8 execution ports per core compared to 6.
It'll also have twice the vector processing performance (thanks to AVX2).

Those are the confirmed features thus far afaik.

That's a great deal then. What cooler do you have (if you have one already)?

Finally got my rig setup, did a quick 4.5ghz OC (3570k) and benched the game. Results are pretty much similar to Venomx1. Still far from your 2500k results

(03-28-2013, 05:54 PM)admin89 Wrote: [ -> ]So Turion X2 is as fast as Llano x4 clock for clock
Turion destroys Llano in single threaded performance and Llano gains 20% performance from background processing (2 cores for Dolphin , 2 cores for...)
Edit : I forgot to include GPU performance . AMD Mobility 5470 is too slow , it does not meet Dolphin minimum requirement -> It bottleneck the CPU -> Turion x2 is faster than Llano x4 in Dolphin after all

Im not sure of the single threaded performance of the turion II vs the llano, neither have an l3 cache and are lower end cpus.
But

Quote: The performance of the processor part depends on the used cores.
Due to Turbo Core, a single core can be clocked up to 2.3 GHz leading to
a performance of a AMD Turion II P540 (2.4 GHz 2MB L2 Cache).

taken from: http://www.notebookcheck.net/AMD-A-Series-A6-3420M-Notebook-Processor.61302.0.html

so at 2.3-2.4ghz for 2 cores of the llano, it says its about equal to the P540 (funny that processor was mentioned with it huh).
So if you overclocked it to run 4 cores at 2.3-2.4ghz, then you should be getting better performance then me, since you have other cores to manage the os at the same time.

Like i said i really optimize my system (tho dont go to extremes), and i did have no extra applications running at the time of the benchmark (i ended all processes that didnt need to be running, like skype and etc. Tho i didnt go crazy and exit services, and windows processes or system processes and etc.
Also i did it after a fresh restart of my computer.

Id say you would probably get better frames if you did a little tweaking to your os, and also did it right after a restart and ahve no other applications running at the same time that dont need to be.



Also the 5470m is above the min requirements

Quote: Average: The ATI HD3650 is not a good choice if you're
looking to run games with high graphic settings. You should expect to
run most games at 1x-1.5x Internal Resolution.

3650: http://www.videocardbenchmark.net/video_lookup.php?gpu=Radeon+HD+3650+AGP&id=585
3650m: http://www.videocardbenchmark.net/video_lookup.php?gpu=Mobility+Radeon+HD+3650&id=487
5470m: http://www.videocardbenchmark.net/video_lookup.php?gpu=Mobility+Radeon+HD+5470&id=58
(not really much in benchmarks for the 3650, but really a pci-e 3650 dekstop card is probably about the same performance as the 5470m)

so id have to say the 5470m should be more then enough for playing at native 1x resolution with no harm on emulation speed. If it is bottlenecking the cpu performance, it cant really be by that much.

But in this benchmark, there is no AA/AF and IR is set to 1x, so having a better gpu isnt going to change this benchmark. So even tho u have a better gpu, it really wont show here.

Please don't use Passmark software , it's inaccurate
Passmark (CPU performance) is even more stupid and the result is messed up completely . Don't ask me why
Dolphin is much different from any PC game or application :
http://forums.dolphin-emu.org/Thread-minimum-specs-video-card
Intel HD 3000 is the minimum requirement for Dolphin ...

Oops , i was wrong (i thought 5470 would have 5GB/sec memory bandwidth) . 5470 does have 12.8GB/sec memory bandwidth (DDR3) and 25.6 GB/sec (GDDR5) -> It does meet the requirement

(03-28-2013, 08:54 PM)admin89 Wrote: [ -> ]Please don't use Passmark software , it's inaccurate
Passmark (CPU performance) is even more stupid and the result is messed up completely . Don't ask me why
Dolphin is much different from any PC game or application :
http://forums.dolphin-emu.org/Thread-minimum-specs-video-card
Intel HD 3000 is the minimum requirement for Dolphin ...

Oops , i was wrong (i thought 5470 would have 5GB/sec memory bandwidth) . 5470 does have 12.8GB/sec memory bandwidth (DDR3) and 25.6 GB/sec (GDDR5) -> It does meet the requirement

well i know computer hardware.... i mean like REALLY know hardware XD

and i dont like passmark either, due to the way they publish benchmark results (they are unfiltered, of all different clock speeds and different types like GDDR3/GDDR5 for the same one, and etc), but its mostly due to the fact, the 3650 is so old, there isnt much else to choose from.

But i do know that the 3650 is about the same performance level of a 5470m, as only 40 less shaders, same amount of TMUs and ROPs, same gddr3 memory, about the same clock speeds, tho the 5470m is at 750mhz vs 725mhz of the reference 3650, and the 5470 memory is clocked at 1800mhz, vs the 1600mhz of the 3650.
And the fact the 5470m is based on 40nm architecture, and the 3650 is based on 55nm architecture, so im pretty sure its better and more efficient architecture.


And i am well aware that dolphin, and other emulation is far different from normal pc software...

Well i now even got 1 more benchmark done
Was able to snag my aunts laptop for a few mins lol


core i5-3210m @2.8-2.9ghz turbo during test
8GB of ddr3 1600 11-11-11-30 memory (dual channel 2x4GB)
intel HD4000 graphics
windows 8 x64
(with integrated graphics they do benefit performance when running faster speed dual channel memory)

Now her computer isnt as optimized, but i did go and exit a bunch of stuff that was running.
But it was not after a fresh restart, as i forgot to do it.

I did have it running on a laptop cooler, and it was clean of dust, so shouldnt have been any overheating (tho pushing cpu to limits while using integrated graphics on the same chip, probably causes the cpu to throttle itself alittle bit id imagine)

FPS 53 / VPS 106 / SPEED 176


FPS 43 / VPS 86 / SPEED 143

FPS 65 / VPS 130 / SPEED 216

May have been able to squeeze a few more fps with a few tweaks, but it does seem that the intel integrated graphics are definitely bottlenecking this cpus performance.

admin89 Wrote:Turion destroys Llano in single threaded performance

Maybe due to higher clock rates since clock rates had to be lowered on llano to save power for the IGP. Their IPC should be nearly identical. They both are based off of the 2nd generation 10H architecture with 1MB of L2 cache per core. Llano might have a slightly IPC lead (5%) due to minor improvements made to the TLB and data prefetcher. Also I highly doubt that having a quad core cpu would boost performance by 20% in this benchmark. So his results seem pretty normal.

jbone1337 Wrote:well i know computer hardware.... i mean like REALLY know hardware XD

Regardless of whether this is true or not please don't tell people this on the internet. This is the equivalent of putting a sign above your head that says "don't take anything I have to say seriously". People use this as a form of validation for their claims as if their opinions about themselves carry some weight. You should not feel the need to advertise this to people if you can back up any claims you make with facts and data.

jbone1337 Wrote:but its mostly due to the fact, the 3650 is so old, there isnt much else to choose from.

This chart is fairly accurate: http://www.tomshardware.com/reviews/gaming-graphics-card-review,3107-7.html

The ranking is based on average performance from tomshardware benchmarks. The 3650 (DDR3 edition) is 3 tiers above the 5470m. The DDR2 edition of the 3650 is 2 tiers above it. And given their specs that makes sense.

If this were another game I might suspect a gpu bottleneck. However given the slow cpu and the fact the WW is extremely light on the GPU (which is one of the reasons that it was chosen) I can safely say that the gpu is not a bottleneck in this instance.

jbone1337 Wrote:and i dont like passmark either, due to the way they publish benchmark results (they are unfiltered, of all different clock speeds and different types like GDDR3/GDDR5 for the same one, and etc)

Our problem with passmark is that it tends to be a poor representation of the performance of most applications, especially dolphin. It is based around multithreaded compute performance (short loops of arithmetic operations with many iterations).

jbone1337 Wrote:But i do know that the 3650 is about the same performance level of a 5470m, as only 40 less shaders, same amount of TMUs and ROPs, same gddr3 memory, about the same clock speeds, tho the 5470m is at 750mhz vs 725mhz of the reference 3650, and the 5470 memory is clocked at 1800mhz, vs the 1600mhz of the 3650.
And the fact the 5470m is based on 40nm architecture, and the 3650 is based on 55nm architecture, so im pretty sure its better and more efficient architecture.

The 5470M has the same memory bandwidth as the DDR2 edition of the 3650 and half the memory bandwidth of the DDR3 edition of the 3650. Both have multiple versions with different memory standards.

The 5470M has 120 GFLOP/s peak shader throughput. The 3650 has 174 GFLOP/s peak shader throughput. Both use the unified vliw5 stream processor architecture (R600 in the case of the 3650m and R800 in the case of the 5470m) so game performance relative to peak performance should be roughly the same for both.

(03-29-2013, 11:29 AM)NaturalViolence Wrote: [ -> ]This chart is fairly accurate: http://www.tomshardware.com/reviews/gaming-graphics-card-review,3107-7.html

The ranking is based on average performance from tomshardware benchmarks. The 3650 (DDR3 edition) is 3 tiers above the 5470m. The DDR2 edition of the 3650 is 2 tiers above it. And given their specs that makes sense.

If this were another game I might suspect a gpu bottleneck. However given the slow cpu and the fact the WW is extremely light on the GPU (which is one of the reasons that it was chosen) I can safely say that the gpu is not a bottleneck in this instance.

jbone1337 Wrote:But i do know that the 3650 is about the same performance level of a 5470m, as only 40 less shaders, same amount of TMUs and ROPs, same gddr3 memory, about the same clock speeds, tho the 5470m is at 750mhz vs 725mhz of the reference 3650, and the 5470 memory is clocked at 1800mhz, vs the 1600mhz of the 3650.
And the fact the 5470m is based on 40nm architecture, and the 3650 is based on 55nm architecture, so im pretty sure its better and more efficient architecture.

The 5470M has the same memory bandwidth as the DDR2 edition of the 3650 and half the memory bandwidth of the DDR3 edition of the 3650. Both have multiple versions with different memory standards.

The 5470M has 120 GFLOP/s peak shader throughput. The 3650 has 174 GFLOP/s peak shader throughput. Both use the unified vliw5 stream processor architecture (R600 in the case of the 3650m and R800 in the case of the 5470m) so game performance relative to peak performance should be roughly the same for both.

Well the 3650 is rated higher then a 5470m, due to the 128bit bus compared to the 64bit bus, so it will have a higher memory bandwidth. And it has 40 more unified shaders, but a slower core/shader clock rate (reference clock), and a slower memory clock.

But in actual gaming performance at a 720P resolution (1280x720/1366x768), they will score about the same (3650 would get maybe 1-2 fps more lol), which that level of difference is negligible.
If you were playing at like 640x480 or something, there would be a larger difference in performance between the 2. But really thats not a real world benchmark when benchmarking a gpu, since no one is ever going to be really playing at that resolution....

Quote:it does seem that the intel integrated graphics are definitely bottlenecking this cpus performance.

I doubt that . Intel HD 4000 is faster than AMD 5470
Both i5 3210M & i5 2450M have same turbo boost 2.9GHz when 2 cores is active . Ivy Bridge is 5-10% faster than Sandy . You should get more FPS with i5 3210M
Try : Open Intel Graphic Control Panel - 3D - Set everything to max