Source: Tom's hardware UK – Keywords: Geforce, Radeon, Budget
Categories: Graphics
MSI Radeon HD 2600 Pro Noise Free Edition
Unlike HIS, MSI ships its card with DDR3 RAM. Nonetheless, the 512 MB of DDR2 RAM (525 MHz) on the HIS card outpace MSI’s 256 MB of DDR3 RAM (700 MHz). The games in our benchmark suite react positively to more video memory, which results in better frame rates at higher resolutions. Activating the Dynamic Overclocking Technology (D.O.T.) in MSI’s driver yields between 2 and 10 percent more performance, closing the gap between the HIS and MSI cards.
MSI cools its Radeon HD 2600 Pro without a fan using a passive heatsink and calls these boards Noise Free Edition. Two heatpipes conduct heat away from the graphics chip and into the cooling fins. The XT version of this card uses three heatpipes. The large cooler is positioned on top of the graphics card, leaving the slot below free for expansion cards. However, your North Bridge cooler should be no taller than 35 mm, or it will come in contact with or even block the card’s large cooler. Since the heatsink is mounted at a slanted angle, one of the heatpipes blocks a memory slot in our test system, as you can see in the image below. Brave souls can bend the cooler a little further upwards to get the offending heatpipe out of the way. However, be very gentle and careful if you do so, as you could easily damage the graphics chip or the copper heatpipe. We were surprised that MSI didn’t think to check the positioning of the cooler for problems in the first place.
The power consumption of the Radeon 2600 Pro is comparable to that of the Geforce 8600 GT. Despite the passive cooling solution, the temperatures remain low. We measured 40°C in 2D mode in Windows and 53°C under 3D load.
3D performance is sufficient to play all current games at maximum quality settings. If you wish to play more complex games like Oblivion or simply want to play at a higher resolution, you can deactivate shadows and FSAA to improve performance. You can make up for this by Turning on HDR rendering or blooming effects which are a real visual enhancement in games such as Oblivion, Rogue Universe or Two Worlds.
You have two choices if you wish to try your hand at overclocking the MSI card. You can either use MSI’s modified driver, which contains the D.O.T. (Dynamic Overclocking Technology) feature, or you can download the official AMD/ATI driver from the web and adjust the clock speeds manually. MSI’s D.O.T. allows overclocking from 2 to 10 percent.
Be warned! Additional overclocking voids your card’s warranty, so any OC-experiments you choose to conduct are at your own risk. For our tests we used the official AMD/ATI Catalyst 7.9 driver. With these drivers, the card doesn’t automatically run at the higher clock speeds that are promised by MSI, as it requires the company’s own modified driver and the D.O.T. feature. Auto Tune, the Catalyst driver’s tool that determines the optimal clock speed, is terminated by VPU recover before the process is completed. ATI has revised the Overdrive section of its drivers, and Auto Tune can now continue where it was interrupted after the computer has rebooted.
However, the results of our Auto Tune attempt were rubbish. On the one hand, the memory was set to run at 685 MHz, slower than the 700 MHz specified by MSI. On the other hand, the GPU frequency was raised from 600 MHz stock speed to a very extreme 795 MHz. Unsurprisingly, the card failed the Test Custom Clock check. We kept testing the same frequencies several more times, but they were failed repeatedly. However, since hardware testers are by necessity more stubborn than any hardware, we kept on trying and finally passed the test.
As it turns out, ATI’s clock speed test is not such a bad idea after all. Our 3D stability benchmark was quickly interrupted by VPU recover, pointing to an overly optimistic overclocking attempt. We finally decided to set the clock speeds manually, figuring that the GPU should be fine at 700 MHz and the memory should be able to handle at least 750 MHz. At these settings, our stability test ran flawlessly. Still, we’re certain that with a little more manual fine tuning (and patience), there is still a little more performance to be squeezed out.
- Previous page HIS Radeon HD 2600 Pro IceQ Turbo
- Next page The Graphics Chips Compared
- The Best Gaming Graphics Cards For Your Money: October 2007
- Can Integrated Graphics Cut It For Gaming Or HTPC?
- ATI's Radeon 2600 XT Remixed
- The Best Gaming Video Cards for your Money: A September 2007 Guide
- The Revolution Will Not be Televised - HD Video on PC
- How Cool Are Thermalright's Graphics Card Coolers?
- The Best Gaming Video Cards for the Money UK: August 2007
- Leadtek's GeForce 8800 Performance Leviathan
- HD 2600 & GeForce 8600: Where's the Mid-Range?
- San Base: Computer Graphics Avant-Garde










How come i started reading a review about entry level cards yet when we get to the benchmarks we have the full Nvidea range on display including a 8800gts.Its just rediculous.The title clearly says DX10 on a budget so how is a 8800gts relevant?
I can understand having the 8600gts in the benchmarks but why no review ?is it by chance that the MSI card isnt really a 8600GTS as it dosent even have a referance chip and the pcb is different?
Even if im wrong on that and you havent used the T2D256E-HD-OC as i suspect you have from the results then why on earth is the 2600XT running 7.6 drivers and the PRO 7.9.
You cant possably hope to conduct a fair and unbiased set of tests if you are not using the same drivers.
There are so many holes in this im wondering if you havent taken the Nvidea dollar as it sems so biased towards them and there products.
Really Really dissapointed
Mactronix
Mactronix