2.6Ghz Intel Core i7 vs 2.8GHz quad-core Intel Core i7 (T4) vs 2.9Ghz quad core Intel Core i7 (3.8)

[fate0311]

The two year old 4890HQ is faster than the new 6700HQ and 6820HQ according to Geekbench 4, CPUBoss, etc. The 4Ghz Turboboost makes a great contribution to many single core tasks while the multi core performance is about the same.

The 6920HQ can match the 4890HQ...but you pay three times less for used MBP with the older chip.

Unless you utilise the extra power of the 450-460 then you won't notice the difference in daily tasks. Even the Intel chip will perform just as well.

Just remember Apple made big claims about the 370X but in real testing it was no faster than the 750M.

The extra bandwidth of the new SSDs are only felt if you can tap into it. Daily use no. But if you copy multi gigabyte files around, yes.

4980HQ not 4890HQ.

You also may want to take a look at the statistics again, because your information is wrong according to this:

https://browser.primatelabs.com/processor-benchmarks

 

[SoyCapitanSoyCapitan]

4980HQ not 4890HQ.

You also may want to take a look at the statistics again, because your information is wrong according to this:

https://browser.primatelabs.com/processor-benchmarks

I benched them myself too. That's why I ended up buying an older MBP this week.

 

[Softwarez]

There's benefits to the new skylake architecture that can't be measured through conventional benchmarks. Skylake's speed shift technology being one of them. It drastically increases how fast the cores can reach turbo states, which decreases latency. It makes the overall system feel much more "responsive" when tasks are bursty in nature, which is basically every situation short of straight up rendering, and prime95 runs.

 

[brynsmith23]

Im not sure if you need to upgrade, that 2013 you already have, should do you fine, thrown in a SSD if you don't already have one and you will be happy with it.

Im upgrading my 15 inch 2.66GHz Duo Core Macbook Pro (Mid 2010) limited to 8gb to an up specced 15inch 2016 2.9GHz with all the upgrades apart from the 512gb SSD.

I will notice that a lot more than you and i can justify it...i've been waiting a long long time for the Skylake to finally come out

I couldn't justify upgrading if i had your laptop, unless you can right it off on your tax as a business expense.

 

[SoyCapitanSoyCapitan]

There's benefits to the new skylake architecture that can't be measured through conventional benchmarks.

I have an OC'd Skylake desktop, a 12 core Mac Pro, and my 4980HQ MBP. Speedstep works on all of them. You just have to run Geekbench on Windows and you can monitor real time clock speed changes with the Task Manager.

Yes, you can measure Skylake performance using conventional benchmarks.

At 4.5Ghz the 6700K gets around 5500 in SC Geekbench.

The X5650 gets a sad SC score around 2500 but is still completely usable with new software and operating systems because it is still sufficiently powerful.

The 4980HQ and the new 6920HQ both have a single core score of around 4500 and multi core score of about 14500. The Skylake has a lower turbo boost and sips less power, but when you take cost into consideration you pay more than double for the new system and will only really feel the benefit if you push a few GPU bound applications.

On the MP forum we also test CPUs and GPUs with Luxmark, Valley, Cinebench, Photoshop, Media Encoder and Tomb Raider. On the Boot Camp side we use the same but also include some newer stuff like Doom's Vulkan mode.

 

[Softwarez]

I have an OC'd Skylake desktop, a 12 core Mac Pro, and my 4980HQ MBP. Speedstep works on all of them. You just have to run Geekbench on Windows and you can monitor real time clock speed changes with the Task Manager.

Yes, you can measure Skylake performance using conventional benchmarks.

At 4.5Ghz the 6700K gets around 5500 in SC Geekbench.

The X5650 gets a sad SC score around 2500 but is still completely usable with new software and operating systems because it is still sufficiently powerful.

The 4980HQ and the new 6920HQ both have a single core score of around 4500 and multi core score of about 14500. The Skylake has a lower turbo boost and sips less power, but when you take cost into consideration you pay more than double for the new system and will only really feel the benefit if you push a few GPU bound applications.

On the MP forum we also test CPUs and GPUs with Luxmark, Valley, Cinebench, Photoshop, Media Encoder and Tomb Raider. On the Boot Camp side we use the same but also include some newer stuff like Doom's Vulkan mode.

Yes, the clock speeds change in real time due to the use of speed step, but the power states are controlled mostly by the OS (or entirely by the OS), which results in response times of around 100ms from lowest P-State to max P-State.

Skylake's speed shift in comparison hands the control of P-States mostly back to the processor, which allows faster, and finer control of frequency. The biggest benefit here is of course efficiency, but with speed shift, min P-State to max P-State only takes 20ms-30ms. That's a pretty significant improvement. (Numbers are taken from Intel's skylake press conference).

As for everything else you've said, I agree 100%. The new processors are barely, if at all better performance wise (when taking into account price) than the old ones. It's more Intel's fault than Apple's though.

 

[SoyCapitanSoyCapitan]

Yes, the clock speeds change in real time due to the use of speed step, but the power states are controlled mostly by the OS (or entirely by the OS), which results in response times of around 100ms from lowest P-State to max P-State.

Skylake's speed shift in comparison hands the control of P-States mostly back to the processor, which allows faster, and finer control of frequency. The biggest benefit here is of course efficiency, but with speed shift, min P-State to max P-State only takes 20ms-30ms. That's a pretty significant improvement. (Numbers are taken from Intel's skylake press conference).

As for everything else you've said, I agree 100%. The new processors are barely, if at all better performance wise (when taking into account price) than the old ones. It's more Intel's fault than Apple's though.

Yes it's Intel's fault but we also have to blame physics