Read more.High frequencies and high speeds. A perfect recipe for the extreme build?
Read more.High frequencies and high speeds. A perfect recipe for the extreme build?
Ooooh, teasers! I do hope that means you'll be testing a number of other DDR3 kits on Trinity - I'd love to see the scaling from 1600 through 1866, 2133 and 2400You may know that AMD is soon to launch a desktop version of its 'Trinity' chip in the coming days. Its IGP component is particularly partial to changes in memory speed, so we'll take another look at this GeIL pack on that processor's launch.
I'm pretty sure the limits you're facing in the graphics based tests is probably due to the IGP being completely maxed out. When you want to test performance of a specific part you need to put it into a system where it is THE limiting factor on performance. So, either pick games and graphics settings where the Intel integrated graphics will have some breathing room or use high end discrete graphics. I understand the purpose is to use a standard benchmark suite, but it's pointless when you're not actually testing the performance of the intended hardware. Every single one of those graphics tests is well within the standard variation of multiple tests on the same hardware.
I think you're missing the point on this one - it's a real-world test intended to demonstrates that for that particular task the speed of your memory makes no real difference. It's quite possible that someone might want to know if faster memory will improve the IGP gaming on their i5/i7-based system, in which case the Hexus IGP benchmark makes a lot of sense (and the answer is "No, save yourself the money and get generic 1600MHz RAM ). Sometimes the point of benchmarking is not to measure the difference between two things, but to demonstrate that the difference between them is insignificant, due to bottlenecks elsewhere. The HD4000 IGP test is a great example of that.
As Tarinder says in the review, the Trinity IGP should be a lot more sensitive to memory bandwidth, and he's going to retest a range of memory with that platform. That *should* show some differences, but again it'll still be an interesting result if it doesn't.
There are currently 1 users browsing this thread. (0 members and 1 guests)