X
    Categories: CPUs

Intel Sandybridge 2500k

Introduction

Well folks, we might be a little late to the party, but we can’t have the CPU section of this site looking as light as it does, so it’s time to add another to that small collective. Today I’m goning to be running the gamer’s best friend, the Sandybridge 2500k, through it’s paces. A few benchmarks, a few games, some beers; lets just all have a good time eh?

 

Specifications

Launch Date Q1’11
Processor Number i5-2500K
# of Cores 4
# of Threads 4
Clock Speed 3.3 GHz
Max Turbo Frequency 3.7 GHz
Intel® Smart Cache 6 MB
Bus/Core Ratio 33
DMI 5 GT/s
Instruction Set 64-bit
Instruction Set Extensions SSE4.1/4.2, AVX
Embedded Options Available No
Lithography 32 nm
Max TDP 95 W
Recommended Channel Price $216.00

Testing and Results

Test Rig

  • CPU: Intel Core i5 2500K
  • Motherboard: Asus P8P67-LE
  • GPU: GeForce GTX 560 Ti 1GB Graphics Card
  • Memory: 4GB PC3-10666 DDR3
  • Storage: 1000GB SATA 7200rpm
  • PSU: 500watt Xigmatek PSU
  • OS: Windows 7 Home Premium 64bit OEM

Methodology

To test CPUs, we put them through several synthetic benchmarks and a few games to see how they perform. Each test is run 3 times with an average score taken in order to avoid any fluke results. The benchmarks run are:

Cinebench
WPrime
3Dmark 11
CustomPC Test Suite
Crysis Warhead
STALKER Clear Sky

For comparison purposes, we used our in-house test system which has the following hardware setup:

  • CPU: Intel Core i5 760
  • Motherboard: Asus P7P55D-E
  • GPU: GeForce GTX 560 Ti 1GB Graphics Card
  • Memory: 4GB Corsair Dominator 1600MHZ
  • Storage: Samsung 5000GB Barracuda 7200
  • PSU: Corsair TX 650w
  • OS: Windows 7 Ultimate x64
 

Cinebench

Cinebench is a real world benchmarking tool that’s based on the Cinema4d engine that’s utilised for 3D content generation. It can be used to compare CPU and GPU performance by rendering a scene of shiny balls and similar objects utilising many modern computing standards.

[easychart type=”horizbar” title=”Cinebench ” groupnames=”i5 2500k, i5 760″ valuenames=”CPU Score” group1values=”5.23″ group2values=”3.87″ ]

 

WPrime

Wprime is a solid multithreaded benchmarking tool used comparing CPU performance by calculating square routes using a “recursive call of Newton’s methond for estimation functions.” Exactly. What he said.

[easychart type=”horizbar” title=”Wprime” groupnames=”i5 2500k, i5 760″ valuenames=”Time (s)” group1values=”11.419″ group2values=”12.675″ ]

 

3Dmark 11

3Dmark11 is the latest 3D based benchmarking tool from Futuremark; this is the benchmark of 3D benchmarks. While it is primarily a GPU performance comparison tool, a graphics card can only perform to its fullets capacity if it’s backed up by a strong CPU.

[easychart type=”horizbar” title=”3Dmark11 Score” groupnames=”i5 2500k, i5 760″ valuenames=”Performance, Extreme” group1values=”4312,1467″ group2values=”4696,1550″ ]

 

CustomPC Test Suite

While it’s a few years old at this point, the CustomPC benchmark suite is nice and easy to use and it offers some good insight into CPU and system performance across several different tests. These include a Handbrake video encoding, GIMP photo manipulation and multitasking.

[easychart type=”horizbar” title=”Custom PC Test Suite” groupnames=”i5 2500k, i5 760″ valuenames=”Image Editing, Videon Encoding, Multitasking, Overall” group1values=”1538,2600,1427,1855″ group2values=”1254,2061,1210,1508″ ]

Gaming

 

Crysis Warhead

Crysis Warhead is still a fine looking game and while most mid to high end systems can run it alright, few get upwards of 60 FPS on the highest settings. To test this we use the FBWH benchmarking tool and run it using the following settings:

Low: Gamer settings, DX10, no AA, 1280×1024
Medium: Enthusiast settings, DX10, no AA, 1280×1024
High: Enthusiast settings, DX10, 8 x AA, 1280×1024
Ultra: Enthusiast settings, DX10, 8 x AA, 1920×1080

[easychart type=”horizbar” title=”Crysis Warhead FPS – i5 2500k” groupnames=”Min, Average, Max” valuenames=”Low, Medium, High, Ultra” group1values=”46.4,28.46,23.3,23.5″ group2values=”62,38.4,31.9,31.9″ group3values=”75.7,42.3,38.1,31.95″ ]

[easychart type=”horizbar” title=”Crysis Warhead FPS – i5 760″ groupnames=”Min, Average, Max” valuenames=”Low, Medium, High, Ultra” group1values=”41.1,40.75,34.6,23.36″ group2values=”80.0,53.3,46,31.6″ group3values=”104.1,65.3,56.1,38.1″ ]

 

STALKER Clear Sky

This is another benchmark from our GPU test suite. Again it’s a nice benchmark to run that’s still reasonably taxing at the higher settings.The ones used were:

Low: Default settings, , no AA, 1280×1024
Medium: Extreme settings, DX10, 1280×1024
High: Extreme settings, Enhanced lighting DX10, 4x AA, 1280×1024
Ultra: Extreme settings, Enhanced lighting DX10, 4x AA, 1920×1080

[easychart type=”horizbar” title=”STALKER Clear Sky – i5 2500k” groupnames=”Day, Night, Rain, Sun” valuenames=”Low, Medium, High, Ultra” group1values=”93.3,78.8,52.6,38.5″ group2values=”132.8,117.6,72.3,49.8″ group3values=”161.1,137.7,80.4,53.6″ group4values=”98,80.8,39,28″ ]

[easychart type=”horizbar” title=”STALKER Clear Sky – i5 760″ groupnames=”Day, Night, Rain, Sun” valuenames=”Low, Medium, High, Ultra” group1values=”85.4,73.9,51.7″ group2values=”118.3,106,70.9″ group3values=”143.2,126.5,77.5″ ]

Power Usage

To test power usage I ran the system at idle for 20 minutes and took the wattage using an wallplug wattage meter. For load results I ran 3Dmark11 for a for 20 minutes and did the same. While CPU, motherboard and RAM were different in each test setup, the GPU was identical.

[easychart type=”horizbar” title=”Power Draw – Wattage” groupnames=”i5 2500k, i5 760″ valuenames=”Idle, Load” group1values=”77,243″ group2values=”130,213″ ]

 

Overclocking

Despite already being quite a workhorse, the Sandybridge 2500k is a strong overclocker too. Pushing it’s performance up is a little different to CPUs of the past, as it’s main clock, while alterable, is linked in with many other controllers. This results in poor system stability after only a few MHZ. While there are methods – including Intel’s own TurboBoost – that allow for some basic overclocking, those with a K based CPU are in luck as their multiplayer is relatively unlocked. This allows for very simple CPU clock altering as some impressive OC results.

After just a short time playing around I was able to push the 2500K from its stock frequency of 3.3GHZ to a much more impressive 4.3GHZ. This was all using the Arctic Cooling Freezer Pro 7 Rev 2.0, hardly the world’s beefiest CPU cooler. With improved cooling it seems likely that I could achieve even higher overclocks with this chip, but considering this PC’s setup I opted not to push it further. Even after only a few minutes of 100% I was already hitting mid 70s on every core which is starting to head in a dangerous direction.

Conclusion

Interestingly while the 2500k walked all over the aging 760 in our synthetic tests, it didn’t faire quite so well in the gaming ones. Probably some naff titles to choose as evidently they arn’t CPU bottlenecked.

Still, this is an impressive chip with some monstrous OC potential. It’s also hardly priced badly 6 months after release as you can get one now for around £120. Hardly too shabby.

Pros

Great performance
Well priced

Cons

Needs P67 chipset for full OC potential

Thanks to Chillblast for sending us this hardware.

Status
Launched
Launch Date
Q1’11
Processor Number
i5-2500K
# of Cores
4
# of Threads
4
Clock Speed
3.3 GHz
Max Turbo Frequency
3.7 GHz
Intel® Smart Cache
6 MB
Bus/Core Ratio
33
DMI
5 GT/s
Instruction Set
64-bit
Instruction Set Extensions
SSE4.1/4.2, AVX
Embedded Options Available
No
Lithography
32 nm
Max TDP
95 W
Recommended Channel Price
$216.00
Whoopty :

View Comments (1)

  • Just want to say your article is as astonishing. The clearness for your post is just spectacular and that i could suppose you are knowledgeable on this subject. Fine together with your permission allow me to clutch your RSS feed to stay updated with impending post. Thanks one million and please continue the gratifying work.