14 DAY TRIAL // Back in March, we found out about a benchmark involving Intel's upcoming Core i7-4770K central processing units, and it wasn't altogether encouraging, so we decided to wait for a second opinion.
Now we have that second opinion, thanks to an "admin" user at RedQuasar forum, but it isn't any more appealing than the first.
In fact, the benchmark results (and there are more of them this time), suggest that the performance of Haswell is more or less the same as that of Ivy Bridge counterparts.
For instance, they are faster when it comes to writing/copying data into memory, but slower when reading from it.
We use the plural because, in addition to the Core i7-4770K that was tested two months ago, the i5-4570 has been benchmarked as well this time.
Another thing is that Ivy Bridge-based 3770K was slightly faster in Super Pi, CineBench 11.5 and Fritz Chess benchmarks.
On the other hand, Haswell has much better graphics performance, with the Cinebench OpenGL test showing a 31% advantage.
Fortunately, the underwhelming results can be explained away, at the moment, through early motherboard BIOS and/or choice of benchmarks.
By the time June rolls around, the mainboard market should have caught up with the new central processors.
If nothing else, programs utilizing new AVX2 and FMA3 instructions will work much better.
A final observation should be made in regard to overclocked states. An overclocked 4770K, even with water cooling, was too hot and unstable to complete Cinebench 11.5, or run intensive benchmarks like Prime95. This problem (instability at high/very high temperatures) carries over from Ivy Bridge, it seems.
To see all the test results, such as they are, one need only click on the image enclosed up on the left-hand side of the article. For a first-hand read, the RedQuasar forum thread can be found here.