GTX 970@1.5Ghz and 2500k vs Q9550 in 70+ benchmarks (+bonus i7-860)

[Azix]

clearly nvidia has a problem with dx11 and needs to improve their drivers.

 

[Maestro1337]

clearly nvidia has a problem with dx11 and needs to improve their drivers.

???

uwotm8

 

[psolord]

Great work touta. The i7 860 should be a lot closer to the 2500k.

For future reference u know what you could do? Since you love to benchmark and stuff turn off HD of i7 860 and test it to compare vs sandy bridge.
would also love to see the 2500k at 4ghz.

Thanks man.

I have finished my 860 benches for a few days now, but new stuff came out and I had to re-arrange my uploads. At this point I think maybe I will wait a little until Witcher 3 comes out and also add Wolfenstein Old Blood, Trine 3 and Killing Floor 2.

Indeed the 860 is doing much better. I think that the 2500k mostly presented better performance due to the extra 800Mhz, but this is mostly due to the high -real life- graphics settings I have used. With lower res the difference could be higher. But as I explained, this whole suite was meant to be focused on the GPU, since I wanted to create a database for future reference.

That being said, a 4GHz 2500k addition would be great, but the amount of work was forbidding. The whole test as it is now, will be based on around 200 runs or so. Same goes for disabling HyperThreading. I would have to do an extra 150 runs to include everything. Not that I am lazy, but it takes more time than I can afford.

I wanted to extract every ounce of performance, from both IPC and Fab Process the chips could give, that's why I clocked them to what they could do.

The good thing is that I record cpu and gpu load and show the graphs at the end of the video, so viewers can speculate of what to expect with lower operating speeds. HT is what it is though. You cannot speculate it off, lol.

Still, at a later point I could add what you suggest, on the most cpu limited tests without vids. Just the numbers.

 

[psolord]

clearly nvidia has a problem with dx11 and needs to improve their drivers.

Sarcasm right?

/Sheldon mode off

 

[Maestro1337]

Touta you can just add what you already have from i7 860 and then later add the rest 4 tests. We are all curious to see the lynnfield in action

 

[Azix]

Sarcasm right?

/Sheldon mode off

yeah. based on seeing other people claim AMD drivers have too much overhead just for losing performance on a dual core chip vs quad i7


A 290x would make this series of benchmarks complete

 

[exar333]

OP - Fantastic work!

One question, subjectively, was there much of a difference in minimums or 'playability' as a whole when you ran these? When the difference was 'noticeable' did it reflect the fps numbers posted, or greater than that might suggest?

Not asking for a bunch more data, but just your overarching thoughts.

 

[psolord]

Ok guys, since I have had some requests to post some i7-860 results already, here is the sum up of the 2500k, Q9550 and Core i7-860, on the most cpu limited benchmarks of the original benchmark suite.

Here is the list sorted by the 2500k vs Q9550 delta, from highest to lowest





and here are the same results, sorted by the 2500k vs core i7-860 delta, from highest to lowest.




As you can see, the 2500k/Q9550 delta dropped from 63% to only 13% for the 2500k/i7-860 comparison.

The clock difference of the 2500k@4.8Ghz and the i7-860@4.0Ghz is 4.8/4=20%, so for 20% more clock we only have 13% more performance. Of course this is due to the fact that this whole benchmark suite, is primarily designed to be GPU limited, with real life, very high settings, hence we have many GPU limits in all runs.

At this point, I believe it would be interesting to redo every 2500k benchmark that has a delta of 20% or more compared to the 860, which would correspond to the operating frequency difference of the two cpus, but this time clocked at 4Ghz, as the i7-860 is.

To stretch things even more, I could also disable the HyperThreading of the i7-860 for another run of these above 20% deltas, so we can see what a Lynnfield can do without the help of hyperthreading, but this is a story for another time I guess!

Of course I will come back to post the complete version of this test, with graphs, bbcode with video links etc.

 

[psolord]

Touta you can just add what you already have from i7 860 and then later add the rest 4 tests. We are all curious to see the lynnfield in action

Done and done!

The normalized clocks results will be posted with results only at a later point. No time for that now.

yeah. based on seeing other people claim AMD drivers have too much overhead just for losing performance on a dual core chip vs quad i7


A 290x would make this series of benches complete

Ok cool.

As for the 290X, this test has indeed awaken many questions.

How a 290X would fare in this benchmark suite and how the deltas would be? If the AMD drivers have such terrible cpu overhead, this test would expose the real life consequence, I believe.

Also how an FX processor would fair, with both the 970 and the 290X?

Unfortunately I cannot afford to invest in such tests right now, especially considering that I am unemployed for a year and half now (what a surprise living in Greece, right). Actually the reason I did this test is because I needed to do something creative and get my mind off things.

If anyone is willing to gift me a 290X though, I will be happy to comply. I will keep the 290X for my trouble though!

The plan from now on is to sell much of my hardware in order to get a 390 (if it is good and reasonably priced) and do a new run of these tests between my 860 and 2500k. The Q9550 is most likely to go.

 

[psolord]

OP - Fantastic work!

One question, subjectively, was there much of a difference in minimums or 'playability' as a whole when you ran these? When the difference was 'noticeable' did it reflect the fps numbers posted, or greater than that might suggest?

Not asking for a bunch more data, but just your overarching thoughts.

Thank you mate.

That is a very good question really, but since the video links are all recorded at 60fps, a viewer can clearly see when something is wrong.

I don't believe in minimums, since they are a split second value. For that reason I have recorded the framerate and frametime graphs of MSI Afterburner at the end of each video, so viewers can have an idea of what is happening regarding both. The fraps result I show also contains minimums, but as I said, this is a split second number.

My benchmarks are very lengthy, so a great number of things can affect the minimums, a HDD access for example, although I tend to avoid those. In some results you could see a 0fps minimum, but still the game would run great!

What I did see however, is that on the Q9550 I perceived some stuttering here and there, which could be from both lower framerate and/or weird frametimes.

One of these occasions was in The Evil Within. This is the run on the Q9550.



The framerate was not bad, but the frametimes were way above 16.6ms

Here is the same run on the 2500k



Much better frametimes as you can see and far more consistent.

In both runs I had 60fps vsync framelock enabled in the options. The game is capped at 60fps anyway, so I wanted to see how smooth or not the two cpus would run it. I did not test it on the 860 however, since the 860's suite, was based on the high deltas between the 2500k and the Q9550.

Now that the problem is better illustrated, I may have to test that as well.

The good thing in all the above, is that as you can see I singled out The Evil Within, because this was the more perceivable, although stuttering occurred in other runs, but it was mostly minor. Meaning most other runs were Ok or consistent with their framerates. Too much data has passed through my head though, so I may have missed something.

Ryse Son of Rome seemed to be a stuttery mess on all systems as well, but I guess this is because the game is too dependent on vsync.

 

[Serandur]

Ok guys, since I have had some requests to post some i7-860 results already, here is the sum up of the 2500k, Q9550 and Core i7-860, on the most cpu limited benchmarks of the original benchmark suite.

Here is the list sorted by the 2500k vs Q9550 delta, from highest to lowest





and here are the same results, sorted by the 2500k vs core i7-860 delta, from highest to lowest.




As you can see, the 2500k/Q9550 delta dropped from 63% to only 13% for the 2500k/i7-860 comparison.

The clock difference of the 2500k@4.8Ghz and the i7-860@4.0Ghz is 4.8/4=20%, so for 20% more clock we only have 13% more performance. Of course this is due to the fact that this whole benchmark suite, is primarily designed to be GPU limited, with real life, very high settings, hence we have many GPU limits in all runs.

At this point, I believe it would be interesting to redo every 2500k benchmark that has a delta of 20% or more compared to the 860, which would correspond to the operating frequency difference of the two cpus, but this time clocked at 4Ghz, as the i7-860 is.

To stretch things even more, I could also disable the HyperThreading of the i7-860 for another run of these above 20% deltas, so we can see what a Lynnfield can do without the help of hyperthreading, but this is a story for another time I guess!

Of course I will come back to post the complete version of this test, with graphs, bbcode with video links etc.

That's awesome, thanks for your hard work.