HASWELL core I5: Performance Preview

[386user]

hi..

i also like reading his posts

i chuckle when i see a model name with the trademark next to it

 

[Azuma Hazuki]

hi..

i also like reading his posts

i chuckle when i see a model name with the trademark next to it

I shake my head. It's one thing to be professional, but quite another to sound like you've been Borg'd.

Christian, unless there's something in your contract that states you need the (R) and (TM) buckybits whenever you post about these things, maybe you should consider not doing it? It makes you sound really corporate, and to a forum full of geeks that is NOT a good thing.

 

[Lepton87]

I havent seen anything on the performance of our next generation of processors

I commented only on this part of his post I didn't suggest that he should go about revealing confidential information etc.I was under the impression that he had access to this internal website called "mark" It just struck me as little odd that he had no desire to learn how well it performs, that's all. If he doesn't, then it changes things.

 

[IntelUser2000]

but even at 1200Mhz it would score ~800pts in GPU subtest while (now old) Trinity gets 1360pts and Richland will score even more more(~10-15%) on stock. Drivers can be tuned though so I suppose 3dmark11 will be higher on the retail product, but it won't come close to Trinity. Also performance in games might be different than these tests so iGPU performance is still unknown.

Actually, based on Tomshardware tests, its a LOT closer: http://www.tomshardware.com/reviews/core-i7-4770k-haswell-performance,3461-6.html

The differences between 4770K's HD 4600 and 5800K's HD 7660D is only about 20%.

The correlation between gaming performance and 3DMark11 is pretty bad on the Trinity parts.

 

[Khato]

http://vr-zone.com/articles/meet-haswell-s-graphics-thump-card-crystalwell-gt3-edram-/19604.html

disappointing if true

But interesting to get an actual die shot! With enough motherboard real estate in the shot to make an educated guess as to the die space. Starting with the larger surface mount tantalum capacitors (appear to be of the 6.0x3.2mm and 7.3x4.3mm varieties) we can extrapolate the likely dimensions for the ICH. Given that the previous mobile ICH had a 25x25mm package dimension the most likely figure for this one given those capacitor sizes seems to be 20x20mm. Using that we can get a slightly more accurate pix/mm of around 17.5, which yields something in the 260mm^2 range (given the quality of the shot I wouldn't put accuracy any closer than +/- 10mm^2.) Then something around 80mm^2 for the memory chip.

 

[Idontcare]

IntelEnthusiast does a fantastic job in my opinion discussing the things he can discuss

+1, Christian is added value around here, I hope he sticks around indefinitely.

 

[Shephard]

So Haswell really sucks. Worse than jump from Sandy Bridge to Ivy Bridge.

 

[Arachnotronic]

So Haswell really sucks. Worse than jump from Sandy Bridge to Ivy Bridge.

Can I please have some of what you're smoking?

 

[Tsavo]

Probably what he cares most about is keeping his job, not losing it by giving out confidential information before the NDA is lifted. (Not sure if you were just joking, but I assume not, so I gave a serious answer.)

Actually I am very impressed by his posts in general, very upfront about his affiliation and gives guarded and carefully thought out answers, not some type of fan-boy pro-Intel comments.

Yep.

Beyond that, I boggle that companies still allow employees to fly the corporate banner on unofficial forums.

IE is totally innocuous, especially now that he quit using that gay non default font in his posts.

JFAMD is a prime example why companies should not allow company representatives to post on unofficial forums.

 

[inf64]

Actually, based on Tomshardware tests, its a LOT closer: http://www.tomshardware.com/reviews/core-i7-4770k-haswell-performance,3461-6.html

The differences between 4770K's HD 4600 and 5800K's HD 7660D is only about 20%.

The correlation between gaming performance and 3DMark11 is pretty bad on the Trinity parts.

That's because Trinity is memory BW limited. Richland will bring more performance in the short term and Kaveri is going to bring considerably more performance in ~6 months.

 

[grimpr]

Actually, based on Tomshardware tests, its a LOT closer: http://www.tomshardware.com/reviews/core-i7-4770k-haswell-performance,3461-6.html

The differences between 4770K's HD 4600 and 5800K's HD 7660D is only about 20%.

The correlation between gaming performance and 3DMark11 is pretty bad on the Trinity parts.

Thats a big fail on the 300$ Core i7 part compared to the 120$ AMD APU and we didnt started taking the driver and image quality situation into more thorough review. Clearly nobody wants to pay more for the crappy Intel igpus on the quadcore line, i personally would welcome an extra 1-2 core for the Intel gpu diespace but Intel wants us to pay for its crappy graphics, i dont buy it.

 

[frozentundra123456]

Thats a big fail on the 300$ Core i7 part compared to the 120$ AMD APU and we didnt started taking the driver and image quality situation into more thorough review. Clearly nobody wants to pay more for the crappy Intel igpus on the quadcore line, i personally would welcome an extra 1-2 core for the Intel gpu diespace but Intel wants us to pay for its crappy graphics, i dont buy it.

Well from looking at the sales figures, apparently people care more about CPU performance and efficiency than they do about igpu performance.

 

[AtenRa]

Well from looking at the sales figures, apparently people care more about CPU performance and efficiency than they do about igpu performance.

99% of people don’t know the difference between the CPU and GPU.

 

[Abwx]

Well from looking at the sales figures, apparently people care more about CPU performance and efficiency than they do about igpu performance.

Because thanks to viral marketing they think that GPU is less
important than CPU...

Ask the guys that did buy laptops with crap integrated gfx ,
the gpu rendered the thing obsolete long before the cpu was
itself oudated , as proved by the DT that kept working well
once the GFX card was upgraded.

Mark my words , SB based laptops with IG will be outdated
long before Llano based ones despite its lower CPU perfs.

 

[aaksheytalwar]

In my experience a new type of CPU is much more than fps in games. It makes games smoother and stutter free even at similar fps.

 

[TakeNoPrisoners]

The only thing that can save Haswell on the desktop is if it overclocks well. If it doesn't....well Sandy Bridge will be the longest lived desktop CPU ever for many enthusiasts.

For laptops it will probably be pretty good, Ivy was fantastic for laptops and Haswell looks like more of the same.

 

[frozentundra123456]

99% of people don’t know the difference between the CPU and GPU.

Never said they didnt.

 

[frozentundra123456]

Because thanks to viral marketing they think that GPU is less
important than CPU...

Ask the guys that did buy laptops with crap integrated gfx ,
the gpu rendered the thing obsolete long before the cpu was
itself oudated , as proved by the DT that kept working well
once the GFX card was upgraded.

Mark my words , SB based laptops with IG will be outdated
long before Llano based ones despite its lower CPU perfs.

Project whatever you want. What percentage of the laptop market do AMD cpus have?

 

[Abwx]

More than you think....
Between 17 and 19% , depending of the quarter , in 2012.

Remember , most laptops have IGPs.

 

[frozentundra123456]

More than you think....
Between 17 and 19% , depending of the quarter , in 2012.

Remember , most laptops have IGPs.

Wow, that's a ringing endorsement.

 

[Abwx]

They started from almost zero a few years back , so it s a good number.

Indeed , they are much downplayed by trolls in a lot of forums who
make their best to hide the fact that globaly Intels products are much
inferiors , the same trolls that once advised to stick to i3s for games
rather than AMDs 4C...

 

[IntelUser2000]

But interesting to get an actual die shot! With enough motherboard real estate in the shot to make an educated guess as to the die space. Starting with the larger surface mount tantalum capacitors (appear to be of the 6.0x3.2mm and 7.3x4.3mm varieties) we can extrapolate the likely dimensions for the ICH. Given that the previous mobile ICH had a 25x25mm package dimension the most likely figure for this one given those capacitor sizes seems to be 20x20mm.

Khato, the 8-series PCH package size is 22mm x 23mm.

 

[Enigmoid]

Thats a big fail on the 300$ Core i7 part compared to the 120$ AMD APU and we didnt started taking the driver and image quality situation into more thorough review. Clearly nobody wants to pay more for the crappy Intel igpus on the quadcore line, i personally would welcome an extra 1-2 core for the Intel gpu diespace but Intel wants us to pay for its crappy graphics, i dont buy it.

?

Other cpu's will have the same igp at very similar clockspeeds.

Anyway, where this really matters is mobile. a10-4600m is about 2/3 the desktop version's power. hd4600 looks like solid competition considering any standard voltage part runs at the same speed as the desktop chips.

 

[Abwx]

?

Other cpu's will have the same igp at very similar clockspeeds.

Anyway, where this really matters is mobile. a10-4600m is about 2/3 the desktop version's power. hd4600 looks like solid competition considering any standard voltage part runs at the same speed as the desktop chips.

The competition is not Trinity but Richland wich will improve
GPU perfs by 20% or so...

http://www.cpu-world.com/news_2013/2013010801_AMD_discloses_performance_of_future_mobile_APUs.html

 

[AtenRa]

The competition is not Trinity but Richland wich will improve
GPU perfs by 20% or so...

http://www.cpu-world.com/news_2013/2013010801_AMD_discloses_performance_of_future_mobile_APUs.html

Even against Trinity at 35W it will not be faster.