Early Skylake Benchmarks only 4% better than Haswell

[superlawyer15]

So ...nearly everyday this forum has had a new thread with questions like "when will the rMBP be updated to Skylake?" "Will it happen at WWDC?" "Will it happen in July?" Will it happen in August?"

Well folks, a benchmark has leaked and Skylake is only pushing out a 4% performance bump over a comparable Haswell chip.

http://www.legitreviews.com/intel-core-i7-6700k-skylake-s-benchmarks-leaked_164456

Yes this is just a leak and we have no way of knowing if it is in fact true, but lets for a second assume it is and lets also assume that Apple ditches the dGPU in a resigned thinner 2016 rMBP; in that case the 2015 rMBP w/ dGPU would be a far more powerful machine overall.

 

[dk808]

Damn, I was hoping to wait it out but if this won't be a huge upgrade I might as well just buy the current MacBook Pro

I'm on mid 2010 mbp and might pull the trigger later this month

 

[Cuniac]

Well the multicore shows an 8% improvement. If this is accurate its not a big jump. Honestly I expect it to be a 15% gain, but only 8% would not surprise me.

 

[Dark Void]

I've been trying to be as helpful as possible on here in the last week but I have to admit the Skylake threads are tiresome.

If you need a laptop now, buy one now. If you don't, wait until you do.

What does it even mean to ask "should I wait?" What will someone that even asks that question be doing on the computer that is so important? Waiting for what is after Skylake after waiting for Skylake?

I digress.

 

[yjchua95]

So ...nearly everyday this forum has had a new thread with questions like "when will the rMBP be updated to Skylake?" "Will it happen at WWDC?" "Will it happen in July?" Will it happen in August?"

Well folks, a benchmark has leaked and Skylake is only pushing out a 4% performance bump over a comparable Haswell chip.

http://www.legitreviews.com/intel-core-i7-6700k-skylake-s-benchmarks-leaked_164456

Yes this is just a leak and we have now way of knowing if it is in fact true, but lets for a second assume it is and lets also assume that Apple ditches the dGPU in a resigned thinner 2016 rMBP; in that case the 2015 rMBP w/ dGPU would be a far more powerful machine overall.

At this rate I'm not interested in the benchmarks - I'm more interested in the technology that Apple will integrate with Skylake (USB-C, TB3, Iris Pro 7200, inductive charging that comes with Skylake...etc). It's no longer about CPU performance anymore.

 

[Codeseven]

At this rate I'm not interested in the benchmarks - I'm more interested in the technology that Apple will integrate with Skylake (USB-C, TB3, Iris Pro 7200, inductive charging that comes with Skylake...etc). It's no longer about CPU performance anymore.

Yup, what he said.

I never got the impression that Skylake was all about superior CPU power, but instead it's about the new generation of technology that it brings with it.

 

[MikhailT]

I'm not sure what you were expecting. Nobody should be expecting any more than 2-4% performance every year, that's not how it works anymore. We've peaked the CPU performance many years ago, which is why we moved toward multi-core since then and then performance per watt and now compute via GPU.

For Skylake, like yjchu95 said, the main reason to wait for Skylake is for the platform changes. TB3 via USB-C, potential DDR4 support, better Iris Pro performance with bigger eDRAM, more perf per watt, and so on.

 

[inhalexhale1]

So ...nearly everyday this forum has had a new thread with questions like "when will the rMBP be updated to Skylake?" "Will it happen at WWDC?" "Will it happen in July?" Will it happen in August?"

Well folks, a benchmark has leaked and Skylake is only pushing out a 4% performance bump over a comparable Haswell chip.

http://www.legitreviews.com/intel-core-i7-6700k-skylake-s-benchmarks-leaked_164456

Yes this is just a leak and we have no way of knowing if it is in fact true, but lets for a second assume it is and lets also assume that Apple ditches the dGPU in a resigned thinner 2016 rMBP; in that case the 2015 rMBP w/ dGPU would be a far more powerful machine overall.

I think most people were hoping/excited for a redesign, and not necessarily some huge % gain with the CPU. I also think it's more of a reaction to Apple not updating the CPU in the 15" model. It's a great computer, but there is 0 reason to upgrade if you have a 2012 or later model.

 

[superlawyer15]

I'm not sure what you were expecting. Nobody should be expecting any more than 2-4% performance every year, that's not how it works anymore. We've peaked the CPU performance many years ago, which is why we moved toward multi-core since then and then performance per watt and now compute via GPU.

For Skylake, like yjchu95 said, the main reason to wait for Skylake is for the platform changes. TB3 via USB-C, potential DDR4 support, better Iris Pro performance with bigger eDRAM, more perf per watt, and so on.

I don't agree with this.
Most people on here in the past few weeks have been making noise about broadwell/skylake CPUs specifically, not the added components that would come with a redesign.

 

[Freida]

So ...nearly everyday this forum has had a new thread with questions like "when will the rMBP be updated to Skylake?" "Will it happen at WWDC?" "Will it happen in July?" Will it happen in August?"

Well folks, a benchmark has leaked and Skylake is only pushing out a 4% performance bump over a comparable Haswell chip.

http://www.legitreviews.com/intel-core-i7-6700k-skylake-s-benchmarks-leaked_164456

Yes this is just a leak and we have no way of knowing if it is in fact true, but lets for a second assume it is and lets also assume that Apple ditches the dGPU in a resigned thinner 2016 rMBP; in that case the 2015 rMBP w/ dGPU would be a far more powerful machine overall.

Skylake is not about CPU. It's what that CPU will bring along that is important.
Thats why Skylake is such a big deal.

 

[MikhailT]

I don't agree with this.
Most people on here in the past few weeks have been making noise about broadwell/skylake CPUs specifically, not the added components that would come with a redesign.

We can agree to disagree but that's the reality, regular CPU performance has peaked and is not going to return to what we saw in the past with big improvements, Intel is hitting the limits of physics faster every year and each time they shrink the process, the more expensive and longer it takes. The only reason to look forward to the next CPU generation is the platform changes, not the CPU itself.

And yet, there are already 4-5 people here who said what I just said, platform changes are what to look for.

 

[superlawyer15]

We can agree to disagree but that's the reality, regular CPU performance has peaked and is not going to change again. The only sole reason to look forward to the next CPU generation is the platform changes, not the CPU itself.

And yet, there are already 4-5 people here who said what I just said, platform changes are what to look for.

I'm not disagreeing with your point as to CPU changes bringing about other additions, but rather to that distinction having been made in the past(with substance) on these forums.

 

[MikhailT]

I'm not disagreeing with your point as to CPU changes bringing about other additions, but rather to that distinction having been made in the past(with substance) on these forums.

Ah. I make no references to what other said on the forum but yea, a lot of people have too much of expectations for future technologies, technology move fast but it is not as fast as we want it.

 

[dusk007]

There certainly were some who heard Skylake - new architecture and hoped for something like Sandy Bridge. But Haswell already taught us that new architecture means little for the CPU. It was already clear for everyone who informed him/herself even a little that Intel made absolutely no promises on the cpu core but tiny optimizations. It is all about platform power, new GPU models (72EU, and 64MB edram) and everything that lives around the CPU on the periphery.
But still Skylake will show quite some difference especially for quad core notebooks in battery life. And performance is more than IPC, a very mature 14nm process will do something for clock speeds as well.
So it is still a lot more significant than just Broadwell.

But ever since Sandy Bridge got the Cache reworked there seems to be little to improve other than bringing down power and putting more cores to work. IPC gains are small and since AMD puts up no competition there is no rush for Intel to add more cores either. If Intel would sell some 8 core with GT2 graphics (would not be bigger than the Iris Pro chips), that would be something, but they don't. With low enough clocks they could.
Maybe AMD with 16nm and Zen can help force Intel to push the limit of 45W again and thow some transistors at something other than the iGPU.

 

[MikhailT]

There certainly were some who heard Skylake - new architecture and hoped for something like Sandy Bridge. But Haswell already taught us that new architecture means little for the CPU. It was already clear for everyone who informed him/herself even a little that Intel made absolutely no promises on the cpu core but tiny optimizations. It is all about platform power, new GPU models (72EU, and 64MB edram) and everything that lives around the CPU on the periphery.
But still Skylake will show quite some difference especially for quad core notebooks in battery life. And performance is more than IPC, a very mature 14nm process will do something for clock speeds as well.
So it is still a lot more significant than just Broadwell.

But ever since Sandy Bridge got the Cache reworked there seems to be little to improve other than bringing down power and putting more cores to work. IPC gains are small and since AMD puts up no competition there is no rush for Intel to add more cores either. If Intel would sell some 8 core with GT2 graphics (would not be bigger than the Iris Pro chips), that would be something, but they don't. With low enough clocks they could.
Maybe AMD with 16nm and Zen can help force Intel to push the limit of 45W again and thow some transistors at something other than the iGPU.

I'm really hoping Zen will be the K6 moment for AMD, there is not much left for them.

 

[christarp]

I'm really hoping Zen will be the K6 moment for AMD, there is not much left for them.

Or the athlon 64 moment..

 

[dusk007]

I am pretty sure they won't level with Intel. They will compete with Cannonlake by the time they get it out most likely. But it might make them much more competitive. If HBM memory is an option, their APUs would certainly teach Intel a lesson. Good enough CPU and better GPU.
Without HBM the GPU side is serverly limited and unlikely to beat Iris Pro with its edram.

They really need Zen and HBM but I don't know if that is feasible at the price point.

 

[maflynn]

Wow, if those numbers hold true, that will be disappointing. I don't have a haswell machine, but I deferred a purchase a laptop thinking that I'd rather not spend a lot of money on a 2 year old chipset.

 

[sebseb]

I took a nano electronics class in college and the professor would explain how nano technology has not had the same advancements as it had couple of years ago. It has been progressing at a really slow pace. Haswell brought a tremendous increase in cpu/gpu bump compared to first gen core i series but you can not expect the same again.

 

[Just Mark]

If I was Intel, I would see Skylake or the promise of Skylake as direct competition with the current chips. They are probably perfectly happy with leaks like this. People on the fence will move forward with purchases of products containing Hazwell and Broadwell processors. Then they release Skylake and it is 10% to 15% higher in CPU performance. Maybe. Certainly we will see big increases in iGPU performance.

 

[superlawyer15]

If I was Intel, I would see Skylake or the promise of Skylake as direct competition with the current chips. They are probably perfectly happy with leaks like this. People on the fence will move forward with purchases of products containing Hazwell and Broadwell processors. Then they release Skylake and it is 10% to 15% higher in CPU performance. Maybe. Certainly we will see big increases in iGPU performance.

I think the leaked benchmarks will hold true, makes no sense for them to allow bad information to spread and then "surprise" everyone with better performance.

As for the iGPU, while I agree that this is where we will see noticeable gains, it's really a moot point if you go the dGPU route.

 

[dusk007]

Leaks like this usually are quite accurate. But again this one only talks about IPC. What happens at equal clock rates. At a mature 14nm process higher clock rates will probably be part of the equation.
It will likely be 15% easily compared to Haswell if not more.

 

[UBS28]

I bought a 2.3 ghz i7 / 16 GB ram /750m / 512 gb SSD back in 2013 when Intel first delayed Broadwell. This is simply the best Macbook Pro ever considering it was still the latest and greatest for almost 2 years and looks to be not much slower in the upcoming years.