Sorry for the multiple posts but for some reason this forum registers the summary as a page break and wont let me post after that point. As you can see I saw a increase in FPS from just having a higher cpu frequency. It was a marginal increase but an increase none the less.freesafety13

I'm surprised at how even it was with two OCX280's. if you get the chance please test BFBC2.

[QUOTE="risc-vs-cisc"]

[QUOTE="hartsickdiscipl"]

The Far Cry 2 Benchmarking tool runs each test 3 times, and gets an average number for Avg. FPS, Min FPS, and Max. FPS. That way you can't just use the 1 "freak" benchmark run as an example and come up with the wrong idea. So put your fears of my testing methodology to rest. There were 3 runs with the FSB/HT overclock setup, and 3 runs with the Mutliplier/CPU clock speed OC. The numbers don't lie.

I don't know where you're getting this "2 fps idea," although a 2 FPS difference on 3 consecutive runs should still be enough to prove that the CPU overclock is more effective. In the results I posted, the difference is 3.4 FPS in the average FPS category, and 3.7 FPS in the Max. FPS category. Even the Min. FPS difference is more than 2 FPS. The fact that all 3 of these numbers were higher with the CPU/multiplier overclock to 3.7ghz is enough to prove my point, not yours. Granted, it's not a big difference in this particular game.. but it is a difference. Also worthy of note- the CPU was running at over stock speed (3.25ghz versus stock 3.2ghz) in the FSB/HT overclock runs, so it even had a little bit of a CPU overclock working for it too! Would you like me to drop the CPU speed to 3.2ghz or below and run the test again? It will just widen the gap.

hartsickdiscipl

So 3.7 VS 3.248 that's 452Hz on the CPU VS 32 on the FSB and you get what 3 FPS difference when the gpu isn't bottleneck and you cant see how you have proved me wright.And the next time somebody tries to use you for ownage like he did speak up instead of attacking me take the time to read through the threads.

My god man.. we're talking about a game that isn't even the hardest on the CPU and it's showing gains by upping the CPU frequency. Did you even look at my 3dmark Vantage results? The overclock of 500mhz that I run everyday is only a 15.6% overclock over the stock speed of my CPU. The FSB overclock of 32mhz over the stock 200mhz bus speed is a 16% overclock. Which one do you think is really harder on the system? I can answer that for you.. the FSB overclock is. Why do I need to read through the threads (which BTW, I did read most of it) when my own results prove what the best solution for myself and other people with like systems is? You asked us to do some testing, so I did. You just don't like the results.

All 3dmark products should not be used or compared with real in game benchmarks AMD,INTEL,ATI and NVIDIA try and mess with the numbers.I And i like your in game numbers they prove my point which was this.

"""""""""I see a lot of bad information going around. That a higher CPU clock will bring higher FPS and people pushing past 3GHz to 3.6 and 4.0 are not understanding whats really going on. So this is the test for yourself and see thread

""""""""""""""""The point of this thread is to show that a high cpu clock can be brought down without changing the FSB and see no performance lost and other things are more important like FSB, GPU,RAM and that each system has it's own sweet spot. Taking A 3.4, 3.6, 4.0 down by only using the multiplier will not have the same stress. So why keep a high CPU multiplier and FSB when you can lower it and remove that extra stress without seeing any performance lost?

""""""""""""That's for you to decide so take the test and find out. This thread is more about finding the sweet spot with ones own systm

""""""""""""Your wright and wrong the point of this thread was to indeed show that once your GPU was bottlenecked your CPU OC was a waste and it still applies today it was also to show that your CPU cant send or receive data faster than the FSB would allow, and to show the importance of a good OC,and to find what works best for you to actually test it and not just up the multiplier and believe it without running benchmarks its also from 9 months ago when people were claiming to get good FPS in crysis by pushing there systems into 3.6 and 4.0 without actually proving it like in this thread where things like this are still said.

""""""""""""I thought my argument was for people to test it for themselves and find out what works best on there system and to learn more about the FSB, CPU, RAM and GPU relationship and to stop people from making ridicules claims based on the I BELIEVE IN THE OC GOD I don't need to test prove or understand anything.Maybe you should read through the thread again ,and your post is full of contradicting statements and baseless accusations,do i really have to dissect this post for you maybe you should read through it again to"""""""""

these are all my quotes which makes me wonder why you think I'm against OC'ing how many times do i have to say this."""""""""""" This thread is more about finding the sweet spot with ones own system"""""""""". which makes me wonder why you became so hostile and started making accusations have you read your posts because it seems like its you who doesn't like the numbers.

[QUOTE="risc-vs-cisc"]

[QUOTE="hartsickdiscipl"]

Far Cry 2 benchies- This time on LOWEST POSSIBLE DX10 detail settings. (640x480, no AA, lowest detail settings)

Phenom II X4@ 3248mhz via a FSB/HT OC to 232/2320mhz, multiplier dropped to 14.0

Avg FPS- 97.74

Max FPS- 146.24

Min FPS- 70.53

And then the results with the FSB at 200, HT link at 2000mhz (stock)

Phenom II X4@ 3700mhz via multiplier upped to 18.5

Avg FPS- 101.17

Max FPS- 149.89

Min FPS- 72.75

Any more discussion needed on this? By dropping resolution and details to the lowest possible DX10 settings, I have taken the GPU bottleneck out of the equation, and shifted the pressure almost completely onto the CPU.. and the CPU at 3.7ghz via simply raising the multiplier beats out the same CPU at just over stock speeds (3.25ghz) with a 16% FSB and HT overclock. Pure CPU speed wins when the GPU bottleneck is taken out the of the equation. This is not surprising.

hartsickdiscipl

Do you realize that 2 FPS is considered the margin of error with every benchmark you see that's why every site runs them multiple times and picks either the highest or lowest scores and 2 FPS is still considered the margin of error and you just proved it works with AMD systems what don't you get you just proved me wright even when your gpu wasn't being bottlenecked, and considering your hostility and baseless accusations constant contradictions I'm staring to wonder if you haven't used a couple FPS to try and make some sort of statement like this "This thread is done.":roll: hartsick= Pure CPU speed wins when the GPU bottleneck is taken out the of the equation. This is not surprising. :|

The Far Cry 2 Benchmarking tool runs each test 3 times, and gets an average number for Avg. FPS, Min FPS, and Max. FPS. That way you can't just use the 1 "freak" benchmark run as an example and come up with the wrong idea. So put your fears of my testing methodology to rest. There were 3 runs with the FSB/HT overclock setup, and 3 runs with the Mutliplier/CPU clock speed OC. The numbers don't lie.

I don't know where you're getting this "2 fps idea," although a 2 FPS difference on 3 consecutive runs should still be enough to prove that the CPU overclock is more effective. In the results I posted, the difference is 3.4 FPS in the average FPS category, and 3.7 FPS in the Max. FPS category. Even the Min. FPS difference is more than 2 FPS. The fact that all 3 of these numbers were higher with the CPU/multiplier overclock to 3.7ghz is enough to prove my point, not yours. Granted, it's not a big difference in this particular game.. but it is a difference. Also worthy of note- the CPU was running at over stock speed (3.25ghz versus stock 3.2ghz) in the FSB/HT overclock runs, so it even had a little bit of a CPU overclock working for it too! Would you like me to drop the CPU speed to 3.2ghz or below and run the test again? It will just widen the gap.

So 3.7 VS 3.248 that's 452Hz on the CPU VS 32 on the FSB and you get what 3 FPS difference when the gpu isn't bottleneck and you cant see how you have proved me wright.And the next time somebody tries to use you for ownage like he did speak up instead of attacking me take the time to read through the threads.

Far Cry 2 benchies- This time on LOWEST POSSIBLE DX10 detail settings. (640x480, no AA, lowest detail settings)

Phenom II X4@ 3248mhz via a FSB/HT OC to 232/2320mhz, multiplier dropped to 14.0

Avg FPS- 97.74

Max FPS- 146.24

Min FPS- 70.53

And then the results with the FSB at 200, HT link at 2000mhz (stock)

Phenom II X4@ 3700mhz via multiplier upped to 18.5

Avg FPS- 101.17

Max FPS- 149.89

Min FPS- 72.75

Any more discussion needed on this? By dropping resolution and details to the lowest possible DX10 settings, I have taken the GPU bottleneck out of the equation, and shifted the pressure almost completely onto the CPU.. and the CPU at 3.7ghz via simply raising the multiplier beats out the same CPU at just over stock speeds (3.25ghz) with a 16% FSB and HT overclock. Pure CPU speed wins when the GPU bottleneck is taken out the of the equation. This is not surprising.

hartsickdiscipl

Do you realize that 2 FPS is considered the margin of error with every benchmark you see that's why every site runs them multiple times and picks either the highest or lowest scores and 2 FPS is still considered the margin of error and you just proved it works with AMD systems what don't you get you just proved me wright even when your gpu wasn't being bottlenecked, and considering your hostility and baseless accusations constant contradictions I'm staring to wonder if you haven't used a couple FPS to try and make some sort of statement like this "This thread is done.":roll: hartsick= Pure CPU speed wins when the GPU bottleneck is taken out the of the equation. This is not surprising. :|

[QUOTE="risc-vs-cisc"]

[QUOTE="V4LENT1NE"] Only thing I have said was the truth to you, why are you running away scared when I bring up the old thread, is it because some of the things you said were so crazy that you are embarrassed? Is it because people are posting benches against you and you cant take the truth? In fact me and a dozen other still dont know what the hell you are arguing because everytime someone trys to discuss against your arguments you get nasty.

hartsickdiscipl

You mean this old thread that ive already linked to and quoted from in this thread and maybe you should take the time to read through this thread and actually comprehend the benchmarks people have posted even hartsick has proved me wright whether he wanted to or not I'm not sure.

"""""but when someone like hartsick does and gives back his opinion you start to get personal and throw out more ridiculous claims.'''''''''''''

I could have sworn it was the other way around.Are you sure your reading through the same thread as me.And I said good day.

I haven't proved you right on anything. I have no idea why you would say that.. unless you're referring to the Far Cry 2 results where my system was GPU-limited at the resolution and detail settings that I ran at. I'll tell you what.. I'm going to crank the resolution and details way down and run the test with near stock speed CPU frequency, but higher FSB and HT. Then I'm going to run it again with my 3.7ghz multiplier overclock. That will take the GPU limitation out of the equation and settle this as far as Far Cry 2 goes.

Have you actually read anything.

hartsick=I haven't proved you right on anything". oh really so in a thread where were talking about pushing a Q6600 past 3GHz to get better FPS and things like this are said """"""""That's pretty bad advice, CPU does have a large impact on framerates."""""""" and this """""""The fastest CPU on the market would still be a bottleneck for any decent cards released in the past 2 years. Go big on the CPU, cool it well, and overclock it. Trust me, your video cards shader output is heavily dependent on the CPU keeping up. Shader output scales with CPU speed increase, even with my old 8800's and Wolfdale..."""""""""""

And hear you are again not even taking the time to read your own posts and contradicting yourself but at least you didn't make baseless accusations this time. """"""hartsick= unless you're referring to the Far Cry 2 results where my system was GPU-limited at the resolution and detail settings that I ran at.""""""""""

[QUOTE="risc-vs-cisc"]

i always had the feeling that high frequencies werent the way to go for me as i saw litterally no performence increase in games. in previous test i found that oc'ing ram mattered alot. even though i didnt go to 3.0ghz its still enough evidence for me to say that high fsb and high mhz ram matters most to me.

[/QUOTE\]

Thanks for taking the test i wish more people would.

Human-after-all

Why did you make it look like i wrote that and he also went with i high FSB why just pick one thing and try to start something.

The problem with most benchmark reviews is they dont say if they have adjusted the QPI manually or if its set to auto. Naturally the QPI will increase as the bclk increases since that has a direct impact on your ram frequency. If I had to guess, I would say that the increase in framerates is more related to the higher ram frequency than the actual cpu bclk or QPI. But the only way to know for sure is to try OC'ing the QPI with the same ram and bclk. Then do it again with a higher ram multiplier as well. I will look into when I get home since I am currently at work and cant do any benching through remote desktop.freesafety13

Thank you I'm very interested in the QPI's performance.

OK, heres my 2 cents worth. With a OC of 3.875Ghz and HT enabled my multi-threaded apps get a huge performance over a 4.0Ghz and HT off. We all know why this is. On the other hand, when benchmarking using 3dMark 06 or Vantage, my 3.875Ghz OC actually scores worse than the 4.0Ghz OC with HT off. That is across the board. My cpu score is higher and my gpu score is higher with HT off and a higher bclk. It stays true to this with Everest memory benching as well. I do not notice any difference in actual gameplay, because, quite frankly, the difference is too small to see with the naked eye.freesafety13

Could you benchmark BFBC2 after seeing these benchmarks I'm curious if its from QPI or is just crossfire.

[QUOTE="risc-vs-cisc"]

[QUOTE="V4LENT1NE"]You say that you want everyone to try it out for themselves but when someone like hartsick does and gives back his opinion you start to get personal and throw out more ridiculous claims. Honestly when me and superclocked were talking to you we weren't even arguing we were just talking and you got all defensive and angry for no reason. We only said in essence that overclocking the CPU increased gaming performance which is fact and you just would not have it and for some reason linked us pointless cpu benchmarks which showed nothing.V4LENT1NE

Absolute nonsense I'm done with you good day.

You mean this old thread that ive already linked to and quoted from in this thread and maybe you should take the time to read through this thread and actually comprehend the benchmarks people have posted even hartsick has proved me wright whether he wanted to or not I'm not sure.

"""""but when someone like hartsick does and gives back his opinion you start to get personal and throw out more ridiculous claims.'''''''''''''

I could have sworn it was the other way around.Are you sure your reading through the same thread as me.And I said good day.

Here's my situation- I have a CPU with an unlocked multiplier. Due to CPU temperatures under heavy load, I can't get a stable OC over 3.7ghz, so that's where my ceiling is whether I get there by upping the multiplier or raising the FSB/HT. My tests have shown me that I don't get an increase in performance over upping my multiplier by using the FSB/HT method. The only thing I DO get are higher temperatures inside my case because I had to raise the HT voltage for stability. So clearly, for me the best option is to use the multiplier to get more CPU speed. And yes.. the CPU speed DOES make a difference in my system in most games that I play.

hartsickdiscipl

Could you test a game that isn't GPU bottlenecked like HL2 I'm interested in seeing how your HT and multi do compared to Intel's FSB.For me higher FSB gave better results than lower FSB and higher multi.