Printable View
i7-4770K was benchamrked in few test at Tomshardware. What I seen, is it about 5-10% better than i7-3770K, no more (without CL HD4xxx optimalization). SO, what do you think? :)
http://www.tomshardware.com/reviews/...ance,3461.html
http://media.bestofmicro.com/R/Y/376...arithmetic.png
http://media.bestofmicro.com/S/0/376...yptography.png
http://media.bestofmicro.com/R/V/376...ginal/lame.png
http://media.bestofmicro.com/R/P/376...al/blender.png
Nice, looks like a decent improvement all around, specially graphics wise. I'm excited and can't wait for oc results and more info.
I mean, for users dont have logic change SB or IB for this...(only for enthusiast). But yes, HD graphic has very nice improvements in performance.
nice scores nice improvement. as you said it doesn't make me upgrade over 3770k but but if i build a system today sure i will wait for this.
For someone building a new PC,if price is right, 4770K is the best you can buy. But for someone who already own SB or IB I doubt this would be a good upgrade. You have to sell your board and CPU ,buy a new combo, and for what? 3-8% IPC improvement with similar OC margin? Seems like a waste of money to me(unless someone doesn't care much for money and wants absolute best). Until AVX2/FMA3 workloads become normal IB at the same clock would be pretty much indistinguishable in legacy workloads from Haswell.
the return performance is less with every generation it seems.Quote:
Originally Posted by informal
time to upgrade from my 2 year and counting 2500k though.
the see3 performance improvement being so small makes me think the only gains in IPC from this processor are coming from code path optimization additions to ss4+.
makes me wonder if all IPC improvements are imaginary now adays and directly related to the new code paths , SSE4 avx.
Not much IPC improvement...
New software will just add an extra code path, old software just gets a minor boost. But it's not imagery if you utilize it. If I remember correctly Sandybridge shuts off part of its SSE engine for SSE3, it's been 2 years since AVX was introduced both companies support it and both console makers have built their next gen machines with it. In new software, SSE3 will be like x87 support within two generations of CPU if x86 and PC's in general survive the 'ARM's race'.Quote:
Originally Posted by Greg83
yes. but what SSE3 perf is showing me.
is the only performance improvement gonna come from this. is in software coded for the new AVX SSE4 additions which finally made use of what was likely available in the processors already. either too hard to code for, with the lazy programmers.
making it appear like this generation is improving on die performance. while its really more optimizing the available then actual rawr performance gains from more compute units.
would these improvements shown , go beyond the software used, or be equal for improvements for all the alternatives.
cause if it only improves the performance with 1% of the software you use. but if you switched all your software, then you could get upto 10%?
ok , due to Bad Thermal Grease in Ivy Bridge , Is Haswell hot as IV ? did Intel fix it ? Hope Intel gives us good new
the culprit behind the overheating issues of Ivy is not the thermal grease but the black adhesiveQuote:
Originally Posted by behrouz
http://i272.photobucket.com/albums/j...edwithNTH1.png
the black adhesive glue is too thick making the gap between the die and IHS wider than it should be.Quote:
Compare all the "a" cases to the "b" cases, the only functional differences in these two test conditions is the CPU TIM. Notice that the Intel stock CPU TIM outperforms the NT-H1 replacement TIM once the CPU-to-IHS gap is identical
(all the "b" cases are ~2-3?C warmer than the "a" cases)
And if we remove the paper shim and drop that IHS down onto the CPU (not perfectly zero of course, there is still some NT-H1 CPU TIM there after all) reducing the gap to as close to zero as possible then we get the "c" cases...and the temperatures show the expected fantastic drops we have all come to expect from delidding our Ivy Bridge chips.
Conclusion: The Intel stock CPU TIM is not the reason Ivy Bridge's run hot, and replacing the Intel stock CPU TIM is not the reason a delidded Ivy Bridge runs so much cooler - the benefits of delidding are entirely due to the resultant reduction in gap height between the CPU silicon die and the underside of the IHS.
Source
Seems like I will grow old with my 2600K. Does Intel really not want people with a Sandy or Ivy Bridge desktop chip to upgrade? Give us at least a 6 core as a minimum already!!! And I don't need silicon spend on an igpu.
GPU with eDRAM on laptop will be a beast. Everything else is as expected..
this. Its been over 6 years since Q6600 was introduced. Its 2013 and still no mainstream 6 cores.. AMD's HSA seems more interesting than this :/Quote:
Originally Posted by nossy23
even haswell mobo's seem more of the same old :( no NGFF slot,No SATA Express, no integrated widi / miracast.. 99% of boards dont come with bluetooh 4.0 or wifi, no decent onboard sound.
and people wonder why PC growth is stagnant. :/
That's an ugly abuse of math, AVX code is up to 100% faster than SSE3 so if you switch you get that boost +10%. And in some cases a simple recompile will get you a significant chunk of that gain. As for lazy coders, optimizations can lead to incorrect output, computers do these equations by brute force, optimized code tends to work by simplifying what you can get away with. But sometimes when your code is solid you have to draw a circle and treat it like a square to speed things up (note exaggeration). Not every good optimization can be figured out in time to ship a product.Quote:
Originally Posted by Greg83
To be honest I think you can't get more SSE3 out of the Intel architecture, every sse port works crazy fast now, Intel and AMD may only have core width left to them as a tool to improve performance in that area.
What I found interesting in THG preview is Sisoft Sandra FMA results. Haswell does perform better than IB/SB but not nearly good enough when you look at the "theoretical" boost FMA could bring- they gained only around 18% of FP performance Vs 256bit AVX codepath. They mention that the FMA results are latency bound so this could be potential bottleneck . This could mean that Haswell-E/EP won't be making huge performance jumps(~2x) in spec fp tests when compared to IB Xeon derivatives. Maybe this is why AMD has not opted to expand their FlexFP unit to native 2x256bit FMA pipes, they expected that intel won't be getting such a massive lead with Haswell's implementation of FMA.
Thanks for the heads up on this article Flank3r , now waiting to see performance and overclockability under cold :) , that will decide if I jump in the bandwagon of Haswell or I keep my good 3770K cpu and Z77 platform.
it seems like my decision to wait out IB and go for Haswell was a good idea
My Q6600, while amazing, is showing its age now.
I think people asking for large leaps in performance are beginning to expect to much. if you think about how much money Intel is putting into the development, it shows that silicon is simply reaching the point where it cannot progress much further.
they've pushed it as far as they can and now it's more down to instruction set optimizations and software optimization to become more efficient (i'm looking at you, Microsoft)
Going from Q6600(OCed) to IB or Haswell is almost like going from Prescott to Conroe ;). Massive leap in overall performance ,massive leap in min. fps in games, massive leap in encoding/rendering. It's money well spent.Quote:
Originally Posted by himaro
Are you a programmer ... probably not because if you were you would understand that the people who are upset the most about not being able to take advantage of an opportunity to optimize are the very programmers you are disparaging. Like Iconyu posted, it takes time ... and we don't get much say in development cycle timing for a project, typically we are given milestone dates and have to meet them. Recoding something that already works is usually last on the list of things to do unless there is a human perceptable different ... so if it takes 1 second, I'm not spending time the boss is not paying me for to get it to happen in 1 ms given that it will only be evident in a synthetic benchmark that has no bearing on the real world.Quote:
with the lazy programmers.
so.. 980x -> 4770k ?
or
980x -> ib-e at the end of the year?
Help?
Hopefully it will oc to 5Ghz on air with decent temps kind of like the 2500k. I also hope there will be a 2c4t chip that can be overclocked to the moon!
Wait for Ivy-E...
Hopefully it overclocks quite a bit better, considering it's built on revised 22nm tech process.Quote:
Originally Posted by rk7p5