i5-2390T vs i3-2130 / i5-2500T

[Xarick]

Hi, i visit this site regularly for bench results and cpu news as i find them very interesting and this is my first post here in the forum.

I was somewhat interested in some bench results from i5-2390T vs i3-2130, however I don't quite understand why the 2390T is priced much higher than i3-2130 or even lower i3-models and performing worse. Is it because of added instructions? What kind of workload would the i5-2390T really shine to justify its price tag?

Also, I would be really interested in a i5-2500T added to the benchmark database, it's another intriguing CPU.

And one more thing, about FX processors, it's quite odd to see them sometimes performing worse than A8-Series in your benchmark, and there's one result for FX-8150 that is very strange, MySQL - Selecting Data with a very low result. Is it really like that?

[gshv]

There are a few reasons why it's more expensive:
- Core i5-2390T has features, that are missing on the Core i3 processors, such as Turbo Boost and Vpro.
- You're paying extra for lower TDP. For "T" processors energy efficiency is a primary feature, and performance is secondary.

2390T should be very good for many types of simple office work (especially when computers are always on), or for use in HTPC, home media players, and similar systems.

Core i5-2500T will be added. in the future This will probably will happen when we get close to Ivy Bridge release.

As for the FX-8150 - yes, it really performs very poor in my MySQL test. I re-tested it a few times. I will try newer version of MySQL when I get some time.

Gennadiy

[gshv]

I think I found that bug:

http://bugs.mysql.com/bug.php?id=15815

I will have to upgrade to newer MySQL version, which means that all current MySQL benchmarks in the database will be invalidated.

Gennadiy

[Xarick]

Thank you for quick reply.

I would like to propose if it's possible at all to add some bench that could focus on the number of cores used instead of either 1 or all cores in use, because there are some different games or programs that still don't make use of all cores available. Also, from a gamer point of view, what type of CPU instructions should I be aware of? What instruction sets do games use? Sorry if it's a newb question.

[gshv]

Thank you for the suggestion about benchmarking, I'll see what I can do.

As for gaming... I don't spend much time playing games, and when I play I don't care about getting the best graphics possible. I play at 1280x1024, and for that resolution my old system with Core 2 Quad Q9550 works just fine. Upgrading from Radeon X1900 to 6950 did make a lot of difference, though If that 4-year old processor is still good enough, then most Intel Core i3/i5/i7 or AMD Phenom II X4/II X6/FX processors will play modern games at good frame rates. In these times having a good video card is much more important for games than having high-performance CPU...

Gennadiy

[Xarick]

Hey again, I was referring about MMX, SSE2/3/4, AES-NI, AVX etc. Do they have any impact in gaming performance? Sometimes I hear in reviews that the CPU is the bottleneck when regarding gaming performance even with top end GPUs (HD 6950+, GTX 570+). What exactly is slowing down?

I noticed in some of your bench results not really matching or being close enough in comparison to my home made results, like for example WinRAR. I got a i5-2500 on a ASUS P8H67-M Pro with 2x 2GB DDR3-1333-CL9 in Dual Channel Mode, and I could get at least 3100 KB/s and if I wait enough, it can go up to 3310 KB/s, but your results show 2719 KB/s at best with ASUS P8P67 and 2 GB dual-channel Corsair TR3X3G1600C9 DDR3. Shouldn't your components wield better results?

PS: Still eagerly awating for some i5-2500T benchs, thanks.

[gshv]

[Xarick]

Hello again. I noticed that different versions of WinRAR produce different results. I am now using WinRAR 4.20 which claims to have severily improved it's algorithm, especially for the 64-bit version.

I can confirm the claim. I used to get 3100 KB/s on an older version, but on this version, I am getting 4900 KB/s. Same system, but with plus 2x 8GB DDR3-1333 CL9 added to it.

[gshv]

Unfortunately, I have to stay with older WinRAR so that all processors in the database are tested using the same WinRAR version. That's one of the reasons why I keep using old benchmarks, even though for some people they seem obsolete.

Gennadiy

[Xarick]

Hi again. I have noticed a flaw in 7-zip benchmark results, or maybe the test was conducted differently than mine.

Perhaps you should provide mention to a summary for each benchmark, about the methodology used and what the test results does focus on.

For example, the 7-zip results I've just seen are about "single-core" efficiency in "multi-threaded" workload if you understand what I mean. I have recently tested in 7-zip a FX-8150 at factory settings with DDR3-1333 CL9 and ASUS Crosshair V Formula, no tweaks or overclocks at all, and I was confused about the disparity in results, later to find out, yours take another approach on the test.

I got 16869 KB/s compressing (670%, 2874 MIPS, 19260 MIPS), 233223 KB/s uncompressing (745%, 2802 MIPS, 20883 MIPS) and a global 708%, 2838 MIPS, 20771 MIPS

The closest results to mine from your FX-8150 is 2465 MIPS(I think?), using DDR3-1866 and ASUS SABERTOOTH 990FX.


Please don't take me wrong, I really do appreciate the work and effort you take in providing all these results, it just seems to be a little bit messed up as to what the results are trying to say.

Thanks again.

[gshv]