Next Generation Emulation banner
1 - 17 of 17 Posts

· From Love and Limerence
6,584 Posts
Discussion Starter · #1 · (Edited)
How do various CPUs perform in PCSX2? How about which CPUs do best clock for clock/at the same frequencies?

Those are the two main questions this will aim to address. This is meant to be a database of sorts, so that being said, the more results we can get, the better. We can use this thread to compile them.

How will we do this? It's simple. We'll use the same version of PCSX2, the same plug-ins, with the same configurations, and the same game(s) at the same spot(s). Basically, we'll attempt to make everything the same with only the CPUs being the differing factors. We'll also use a frequency range that is large enough to compare different CPU types, but limited enough attempt to get more than one result of a the same speed to compare the different CPUs at the same speed. Note that there will always be variance across systems, so we'll assume anything within about a 3% difference is within margin of error.

Now, who has hundreds of CPUs to test, and the time to test them all at various clock speeds? Not me, and not many of us, so the point of this is to let many users add their results. So that being said, as long as you have a decent control over your CPUs clock speed (or if it just happens to already be at one of the speeds being tested), then you're more than welcome to add your results. Just look up your CPU, and make sure it qualifies (or can be adjusted to qualify for) one of the listed frequencies. You can, of course, submit multiple CPUs, or even the same CPU at different frequencies.

Without further ado, here's the list. This is my starting point. Feel free to suggest more. For now, I'm only including desktop CPUs, and not laptop or server CPUs.

CPU Chart

Celeron, Pentium 4, Pentium 4 Extreme Edition, and Pentium D (Netburst Architecture)
Note whether or not Hyper-threading is on or off where applicable with these CPUs.

Any Intel Celeron (socket 478 Northwood-based) @2.6GHz, @2.8GHz, @3.0GHz
Any Intel Celeron D (socket 478 Prescott-based) @2.6GHz, @2.8GHz, @3.0GHz (this is not a dual core CPU)
Any Intel Pentium 4 (Northwood) @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz, @3.4GHz, @3.6GHz
Any Intel Pentium 4 (Prescott) @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz, @3.4GHz, @3.6GHz, @3.8GHz
Any Intel Pentium 4 Extreme Edition @3.2GHz, @3.4GHz, @3.6GHz, @3.8GHz
Any Intel Pentium D @2.8GHz, @3.0GHz, @3.2GHz, @3.4GHz, @3.6GHz

Celeron, Celeron Dual-core, Pentium Dual-core, Core 2 Duo, and Core 2 Quad (Core Architecture)
Extreme Edition versions of these CPUs are fine, as they're the same thing performance wise as their standard versions at any given clock speed.

Any Intel Celeron L (Celeron 420, Celeron 430, Celeron 440, etc.) @2.6GHz, @2.8GHz, @3.0GHz
Any Intel Celeron Dual-core @2.6GHz, @2.8GHz, @3.0GHz
Any Intel Pentium Dual-core (E2xx0 class, Allendale, 1MB L2 cache) @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz, @3.4GHz
Any Intel Pentium Dual-core (E5x00 class, Wolfdale, 2MB L2 cache) @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz, @3.4GHz, @3.6GHz, @3.8GHz
Any Intel Core 2 Duo (E4x00 class, Allendale, 2MB cache) @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz, @3.4GHz
Any Intel Core 2 Duo (E6xx0 class*, Conroe, 4MB cache) @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz, @3.4GHz
*Note that the E6300 and E6400 should be included among the E4x00 Alendale class, as that's what they really are.
Any Intel Core 2 Duo (E7x00 class, Wolfdale, 3MB cache) @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz, @3.4GHz, @3.6GHz, @3.8GHz, @4.0GHz
Any Intel Core 2 Duo (E8x00 class, Wolfdale, 6MB cache) @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz, @3.4GHz, @3.6GHz, @3.8GHz, @4.0GHz
Any Intel Core 2 Quad (Q6x00 class, Kentsfield, 2x4MB cache) @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz, @3.4GHz, @3.6GHz
Any Intel Core 2 Quad (Q8x00 class, Yorkfield, 2x2MB cache) @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz, @3.4GHz, @3.6GHz
Any Intel Core 2 Quad (Q9x00 class, Yorkfield, 2x6MB cache) @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz, @3.4GHz, @3.6GHz, @3.8GHz, @4.0GHz

Core i7 (Nehalem Architecture)

Any Intel Core i7 (9x0 class, Bloomfield) @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz, @3.4GHz, @3.6GHz, @3.8GHz, @4.0GHz
Any Intel Core i7 (8x0 class, Lynnfield) @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz, @3.4GHz, @3.6GHz, @3.8GHz, @4.0GHz

Sempron, Athlon 64, Athlon FX, and Athlon X2 (K-8 Architecture)

Any AMD Sempron @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz
Any AMD Athlon 64 @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz
Any AMD Athlon FX @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz
Any AMD Athlon X2 @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz

Phenom, Phenom II, & Athlon X2 (K-10 Architecture)

Any AMD Athlon X2 (Kuma) @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz, @3.4GHz, @3.6GHz, @3.8GHz, @4.0GHz
Any AMD Phenom X3 @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz
Any AMD Phenom X4 @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz
Any AMD Phenom II X3 @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz, @3.4GHz, @3.6GHz, @3.8GHz, @4.0GHz
Any AMD Phenom II X4 @2.6GHz, @2.8GHz, @3.0GHz, @3.2GHz, @3.4GHz, @3.6GHz, @3.8GHz, @4.0GHz

I specifically have a feeling my knowledge of AMD Athlon K-8 is lacking, as the Athlon 64s and Athlon X2s undoubtedly have different classes that need breaking down, since I presume there are performance differences within there. I'll let someone else fill me in there.


Now, without further ado, here's the setup you must use. Remember that it doesn't matter if it's not the latest beta, as the only thing that matters is that everyone uses the same version so we get more consistent results. Remember, we're not aiming for the most speed, but rather the comparison of the CPUs.

First of all, we'll be using PCSX2 0.9.6, so grab that, which includes all of the needed plug-ins, and install it.


CPU Config: Configure it as follows (the Multi-threaded GS Mode should only be enabled on dual core, quad core, or Hyper-threading CPUs).

Patches: Leave all of them disabled/don't use any.

Gamefixes: Leave all of them disabled/don't use any.

Speed Hacks: Leave them all disabled/don't use any (the esc hack doesn't matter if it's used)/default.

Advanced: Leave them all default.


Graphics: GSdx (version applicable to the instruction set of your CPU).

You may be thinking "why not have everyone use the same version, like, say the SSE2 version?", but since the extra instructions sets are technically a benefit of the better CPUs, they should be able to use them. So long as it's GSdx and configured as is listed, it can be the SSE2, SSSE3, or SSE4.1 version.

We're restricting it to DirectX 9 mode, since the difference between DirectX 9 and DirectX 10 (the latter is faster) falls outside of the CPUs themselves and is dependent on the OS and GPU.

For the same reason, we're using Shader Model 2. It's GPU dependent, and we're after CPU results.

Make sure "Native" is checked.

Make sure "Texture filtering" is not checked.

Make sure the rest of your check boxes are either unchecked, checked, or faded checked to match the image.

Configure it as follows.

Sound: ZeroSPU2

Configure it so the top check box, the "Time scaling" mode, is checked, and nothing else.


Make sure you're running from an ISO to prevent slowdowns from disc reading skewing the results.

It needs no configuring outside selecting the ISO location.

None of the rest of the plug-ins should matter, but I could very well be wrong. Just leave the rest at default (besides the pad plug-in and configuration, which won't matter for performance).


I'm going to try and keep the scenes near the beginning of the games, ala, so you would be able to start it up without a save and get there in not too long.

This section, as of first creating this thread, is obviously incomplete. Until I get game suggestions, I'll stick with the one common one.

Final Fantasy X

I decided on these five scenes.

The title screen.

The very first scene in which you have player control.

Standing amidst the fans some seconds later (stand as close to those lines as I did as possible).

Naming the main character.

The first battle scene.

That's about it, until I get more game ideas.

Feel free to start posting up results (screenshots are needed) and giving suggestions for games/scenes. It's a good idea to list your basic system specifications (CPU, GPU, RAM, and perhaps OS). I wager we'll have a few with GPUs that hold them back regardless of the CPU limited settings.

If I overlooked any glaring or obvious mistakes, go ahead and mention them. I'm sort of wondering if this would be/would've been better done at, to be able to get more results, and since it's now the official forums?

Also, my own results aren't included yet. I'll of course do that myself, but for now, I'm getting the thread started.

· Registered
437 Posts
So we are going to use the latest official version which is pcsx2 0.9.6. There is a problem though. The version that is shown in the picture configurations you posted have different options than the version you posted for use. For example in the gsdx version available through that download, it gives us the option for Pixel shader 2.0 or 30. or to use the NLOOP hack. Which one options should I use?

I think that you can do the test at and here. Over there I had asked a question about how the i7 measures up to the core 2, they said that they know that it is faster, but they don't know by how much. They said that it would be wonderful if someone could test things out for them.

· From Love and Limerence
6,584 Posts
Discussion Starter · #6 · (Edited)
I did alot of copying and pasting with the CPUs for the formatting (and for the Core 2, I copied it all from the E2xx0 line), so I forgot to change that to Wolfdale. I'll fix this.

The Phenom II X2 is included, but as it is called the Athlon X2 (Kuma), that is what I have it down as.

redlofredlof, you have some sort of AMD Athlon X2 if memory serves me, right? How does that not fall under "Any AMD Athlon X2", or does yours simply run slower than 2.6GHz? I'll add more speed ranges if there's reason to, but it'd be worthless to add, say, 2.4GHz for the AMD Athlon X2 unless I did for other CPUs, as the purpose is to compare them. Should I add 2.0GHz, 2.2GHz, and 2.4GHz to the ranges of many of the CPUs?

Also, should I have included the K10-based Sempron LE, or even the K7-based Athlon XP? I thought the Athlon II was an upcoming CPU?

Lastly, I was hoping the screenshots would be largely the same. I didn't actually download the official 0.9.6 version myself yet, so I'll do that now and then update the screenshot.

I guess I'll also post this over at

· Registered
437 Posts
I would show my results if I had FFX. I'll wait for the next game to be tested.

The Phenom II X2 is included, but as it is called the Athlon X2 (Kuma), that is what I have it down as.
Nah. There is an actual Phenom 2 X2 that is not named Athlon. Same with the Athlon II. The Kuma line seems to be discontinued since I'm looking on newegg and there are none available for order.

AMD Phenom II X2 550 and AMD Athlon II X2 250 Processors Review - X-bit labs

I thought the Athlon II was an upcoming CPU?
It was already released a few months ago. the first is at 3.0GHz and the newer versions are 2.9GHz and 2.8GHz. They are all priced at sub 100 price tags. and they actually perform decently. They are excellent bang for buck wise.

· From Love and Limerence
6,584 Posts
Discussion Starter · #9 ·
Okay, it appears the AMD Athlon X2 (Kuma) is really either the AMD Phenom II X2 or the AMD Athlon II X2 (from the article, I'm guessing the former). Another member has one and claimed it was simply called the AMD Athlon X2, so that's what I went with. I don't know what's what with AMD anymore. There's a Phenom X4, a Phenom X3, a Phenom II X4, a Phenom II X3, a Phenom II X2, An Athlon X2, and an Athlon II X2. Wow, how confusing. Next we'll see an Athlon II X4? Can someone with alot of AMD knowledge help in this?

It's perhaps a good thing I can't register on yet (I never got the confirmation e-mail, and it's been a few hours)? This thread isn't ready just yet apparently.

Phil, could you do the test at either 3.8GHz or 4.0GHz? I'm trying to do x.xGHz numbers that have the second digit an even number (ending in 0, 2, 4, 6, or 8) to make it simple.

I'm also aware some CPUs have some "special" default speeds, like 2.26GHz, 2.66GHz, 2.83GHz, 3.16GHz, 3.33GHz, etc., but it'd be too sporadic to include them all. By starting at ~2.6GHz (may lower it to 2.0GHz), and going up in 200MHz increments, there's a better chance of getting results that are at the same speed as another CPU. The overall goal is to compile a database of comparisons, not just compile a table of all the results possible. Remember, we're not going for highest speeds. We're going for the most results we can get, so if anyone wants to test their CPU at all the listed speeds for their "class", that'd really help.

I should also mention that it's a good idea to list your basic system specifications (CPU, GPU, RAM, and perhaps OS). I wager we'll have a few with GPUs that hold them back regardless of the CPU limited settings. I'll edit this part into the first post.

· From Love and Limerence
6,584 Posts
Discussion Starter · #11 ·
Okay, so so far, we have two users with with Core 2 Duo Wolfdale CPUs to test, and two users with Core i7 9x0 CPUs to test.

bkwegoharder, go ahead and recommend some games. I only included Final Fantasy X as I know it's the most common.

· From Love and Limerence
6,584 Posts
Discussion Starter · #13 ·
Of those, I have Dragonball Z Budokai Tenkaichi 3 and Metal Gear Solid 2. That's it. I had the three main Grand Theft Auto games, but my brother traded them in (many of the Playstation 2 games were both of ours, and after a while, we sort of split them based on who played which ones more). Of course, this isn't just limited to games I have, but as I know you wanted to compare your CPU to a Core 2, I'll add Dragonball Z Budokai Tenkaichi 3 for the next one.

· The Master of Disguise
811 Posts
Well, it took some delay from me but here are my results using a E8200 @ 3.2GHz, everything else of my computer specs can be found below.

Screenshots of the results:

I'll try to post some more results with different clock speed later.

· Gyrobyte
117 Posts
Q6600 @ 3.2Ghz, HD4770 OC'ed to ~HD4870 level, 4GB ram, Vista Sp1.

1 - 17 of 17 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.