Is i7 2600 just a hype?

timsu

Distinguished
Apr 20, 2011
3
0
18,510
I am hoping to get some input from you all. I've been waiting and waiting to play around with a 2nd Gen Intel CPU machine and finally got to do so today. It is a Lenovo m91p with i7 Core 2600. This machine has 4GB of RAM, runs Win7x64 and has a 10K WD Velociraptor HDD.

I use Photoshop Lightroom 3 and Photoshop Extended CS5 mostly and I have play with numerous machines in the past and I actually came up with my own way of timing a certain task to give me an idea how fast this "same task" runs on various machines.

To my utter disappointment, this i7 Core 2600 machine runs slower than a i7 870!!! By a lot! Have any of you out there experienced this? I am very puzzled. All the hype about how fast Sandy Bridge is.... and now I am not even sure if I want it.

Thanks in advance for your comments and advice!
 

Umbongo

Distinguished
Sep 24, 2008
51
0
18,640


Try running other benchmarks like SPEC CPU tests, Cinebench, Geekbench. If they show the system performing worse than your i7 870 then something is wrong with the CPU (unlikely), system configuration or another component. Otherwise it is your test method. A Core i7 2600 running correctly is superior to a Core i7 870 running at 2.93GHz.
 
http://www.tomshardware.com/reviews/sandy-bridge-core-i7-2600k-core-i5-2500k,2833-15.html If my math is correct that's 10% difference in PS but overclocking helps a lot. I've seen in other benchmarks where the 2600k is twice as fast in photoshop as the 870 when overclocked. Photoshop is a program that can be bottlenecked by any component of your system. Depending on what you're doing hardware acceleration will also be affecting performance. What are the rest of the specs and is everything running to spec?
 
I looked at the site and configured a machine, and a 10k hard drive wasn't one of the options. Unless the owner added their own drive, it was using a small (only up to 500GB) 7200RPM one.

Was the Lenovo's 4GB of RAM single-channel or dual-channel? The ordering system offers both, and the single-channel option is first.
Was the Lenovo using integrated graphics? Even the discrete graphics card they offer sucks.
Was the competing 870 machine configured anything like the Lenovo hardware-wise?

Most OEMs like that configure their machines for reliability and compatibility rather than speed. It wouldn't surprise me if they turned off performance options in the BIOS and software so they would have fewer problems.

All of the above could certainly have resulted in the performance discrepancy. I would certainly not expect two similarly-configured machines to perform the way you say, timsu. The 2600 should win just about every comparison, whether by a small or large margin.
 
The processor is just a portion of the overall puzzle.

Additionally, the 2600 utilizes SSE4.2 (and AVX instructions) in a much more efficient manner than processors from earlier generations. In fact, some software ran worse with SSE4 in earlier generations but it was used for comparisons because AMD processors ran even more worse with it enabled and showed a much larger discrepancy with Intel. In your testing you may have found one of the "boogers" that reveal themselves from time-to-time.

RAMs, disk I/O and the motherboard chipset also play a large factor beyond instruction sets.

You should have no worries -- unless your testing is a large example of your overall workflow. As software is further optimized (in AVX as an example) the 2600 should perform on par, and exceed that, of the 870.


 
Something seems to be slowing you down, because my work computer with the Intel® Core™ I7 2600K blows the socks off my home system with an Intel Core I7 860. I have a guy in the office with a Intel Core I7 950 and I am still out performing him with the Intel Core I7 2600K by a good bit.

Christian Wood
Intel Enthusiast Team
 

timsu

Distinguished
Apr 20, 2011
3
0
18,510
Hi everyone, thanks for all your help and comments. I really appreciate it!

After a few sleepless nights and thinking what's wrong... I finally realized that the test I used on this i7 2600 was not the same as the previous ones I used!!! (Of course, of all the test I had to make mistake on, it had to be "THIS ONE") I kept clicking a different preset and thus the results will obviously be different. After I used the "same exact" test and ran it again, the i7 2600 did out perform i7 860. BUT not by a lot. So the result was still a bit disappointing.

I understand in a strict testing environment, you would definitly want to make everything as consistent as possible (RAM, HDD, Graphics Card...), but in the real world, unless you have tons of money or you work at Intel or something, it's kind of hard to do that. So I try to record down the variables each time I test. But mostly, they are all 7200K HDDs, 4-8GB of RAM, all running Windows 7 64 bit, all running Lightroom 64 bit. My purpose is just to find out roughly how fast is the latest consumer computer doing the work that I usually do in Lightroom.

The fastest computer I've tested this year was a Dual Xeon X5650 ThinkStation with 16GB or RAM. There is only one minor issue, it costed $6000. Needless to say, I would rather wait a few more seconds/minutes to do my work on a more average high-end PC and use the money somewhere else.

Once again, thanks for the comments and suggestions.