This article will test the new processors Intel Haswell, whose announcement was held in the early summer of 2013:
As their opponents have chosen the following models:
In the first article, every game except skyrim seemed to be gpu limited at 1080p.
But still, disappointing results for Haswell in gaming, based on the two articles.
I do agree that at current prices, if one can find Ivy Bridge cheaper than Haswell, which it usually is, Ivy would be my choice for gaming. If one uses a lot of productivity apps where Haswell shows more improvement, then it could make sense.
Also interesting is that hyperthreading seems to make little difference for quad cores, and the i3 is still quite competitive, based on the second article.
Hardware.fr has a test of various Haswell SKUs versus their IVB/SNB predecessors. No performance regressions, although the improvements aren't breathtaking either. But it's enough to see that the ocaholic benches seem to be useless.That's why I asked if there any other articles testing the cpu's in a bunch of games. One bench of skyrim isn't really enough to prove the ocaholic benches are bogus.
Well, if Haswell scores worse than IB/SB, there must be some error, certainly. However, the point is, except in a few games, its performance is about the same as its predecessor. Which isn't exactly news, right?!Hardware.fr has a test of various Haswell SKUs versus their IVB/SNB predecessors. No performance regressions, although the improvements aren't breathtaking either. But it's enough to see that the ocaholic benches seem to be useless.
The ocaholic one seems to be very "unique".
better?What kind of a dumb setting is this? Benching skyrim calls for ultra settings (especially shadows). Skyrim is noticeably less demanding for the CPU on low settings.
Fact of the matter is that anything on that chart is easily playable at more than 60 fps.
max settings, other than AA, should always be used when testing cpus because many settings impact the cpu too. I can lower my res to 1280x720 in some games and be cpu limited even with my 2500k at 4.4. I can hit a wall way before 100 fps in many cases thus proving that all the gpu power in the world will not let you take full advantage of 120/144 hz screens.It doesn't make a difference. Low settings at 720p = cpu limited. High settings at 720p = cpu limited. This, being a CPU benchmark, the only goal is to show the strengths and weaknesses of the CPUs in question. Low or high does not matter. The only ultimate goal is to remove the GPU from the equation and that's that. The benchmark did just that despite the "low" settings.
Low or high also will not change the results. Most games will favor the 4770k in ST performance in single screen resolutions.
I am talking about max settings (other than AA) and using a lower resolution.Max settings may stress the CPU more, but if their net effect is to mask the differences between CPUs in a CPU test, then to do so is counterproductive.
It might be best to find the highest graphical settings within a given test setup that still illustrate significant differences between CPUs. This point takes time to find, though.
It doesn't make a difference. Low settings at 720p = cpu limited. High settings at 720p = cpu limited. This, being a CPU benchmark, the only goal is to show the strengths and weaknesses of the CPUs in question. Low or high does not matter. The only ultimate goal is to remove the GPU from the equation and that's that. The benchmark did just that despite the "low" settings.
Low or high also will not change the results. Most games will favor the 4770k in ST performance in single screen resolutions.
Yes but the point is to determine real world capabilities too.
you said low or high does not matter and that is WRONG. again many settings impact the cpu so using low settings is stupid as it tells you nothing about how the cpu can handle the game. you can run gta 4 smoothly on low settings with a low end core 2 duo but on max settings even a 2500k cant stop the game from crapping itself in areas. so the way to test cpus is to run them on all high settings except for AA and use a low res.Are you serious? It isn't the determine real world capabilities. The readership there and at desktop oriented websites do not care about 720p "real world capabilities". This isn't a laptop benchmark-athon, give me a break man. This is a desktop CPU benchmark and 99.9% of such users at the readership of HWC have discrete GPUs.
It is to determine the CPU performance with the GPU performance removed from the picture. Period. Real world capabilities? Give me a break. That isn't what the readership there or the readership at most Desktop PC oriented websites are concerned about with CPU benchmarks. They removed the GPU from the equation and that's that.
Your vision is distorted by some notebook benchmarks. Owners of the 4770k that play PC games do not give a crap about 720p gaming thus the notion of "real world capabilities" is laughable. It is not a notebook oriented website - that methodology makes SENSE on a little laptop but doesn't make sense for a desktop. Do you own a desktop? Don't apply your notions of laptop benchmarking to a desktop oriented website like HWC, Guru3d, among many others. The only goal was to remove the GPU from the picture. 4770k desktop owners that play PC games own discrete GPUs and DO NOT care about 720p "real world capabilities".