Intel Skylake / Kaby Lake

[Arachnotronic]

Ah, I've forgotten what exactly Kaby Lake entails. Was Skylake not Gen 9? Regardless, its IGP was definitely a letdown.

Skylake was Gen 9, Kaby Lake looks like tweaked Gen 9 with much better codec support. Marketing slides also say "improved core performance" so I think CPU architecture gets tweaked/enhanced as well.

 

[frozentundra123456]

I actually was fairly pleased with Skylake igpu, although with all the time they have had and a big process node lead, I think their mainstream should be able to match highest end Kaveri. But at the with high end cpus, who really cares about the igpu anyway?

 

[AtenRa]

How long would they be waiting for?

Should be released in Q2 2016, that is 4 to 6 months tops from today.

 

[Fjodor2001]

Thanks! I think there is zero likelihood that I'd buy a Zen CPU unless it crushed Intel in someway (like the price/performance of the K7 and outright performance of the K8) with decent perf/watt numbers. I would be delighted to see Zen match the performance of current Intel CPUs, but I think it's unlikely and the best quality motherboards seem to be Intel based.

I'm at the point where I want to upgrade, but coming up with the dosh is a problem since I like the longevity I've gotten out of an old high performance system. I'll probably wait till I can afford a 16FF+ next gen GTX 980 class GPU, 6-8 core CPU, 32 GB of 2400 MHz DDR4 or faster and at least a 1/2 GB M.2x4 NVme SSD all on a solid motherboard w/good overclocking.

I'm looking for something similar: 8 cores, 16/32 GB DDR4, M.2x4 NVMe 1 TB SSD, 4K 16:10 monitor (IPS or OLED). Not worth upgrading for me until that is available at reasonable cost.

As for Zen, you can be pretty sure it will no beat the latest Intel HEDT CPU in pure performance. But given it's a completely new uArch, 14 nm and 8 cores, chances are it'll come pretty close. So if it's only 10-20% behind in performance, wouldn't it be an option if it's priced noticeably lower? And if anything it'll bring back some competition to the high-end desktop segment, potentially forcing Intel's prices lower too. That's how I reason.

Also, the motherboard will likely be much cheaper for Zen as well since Zen is compatible with the new AM4 socket that will be used for AMD's mainstream APU's next year too. AM4 is upgraded to the latest I/O standards, which is important to me. I don't need the other stuff on the Intel HEDT LGA2011 socket motherboards like quad channel DDR4, so that is just a waste of money.

 

[dark zero]

If is only 10% performance disadvantage, I'll think about that. But if is more than 15%, unless cost 35% less than the competition, is a big nope

 

[Magic Carpet]

Just noticed that 6700K's Tcase temp is lower by some 10C compared to 4790K. If anything 2600K has a similar Tcase to 4790K. In fact, Skylake seems to be the first Intel chips in years with Tcase below the 70C mark. The last one was a 45nm i7 920 D0. What gives?

6th Generation 14 nanometer: i7 6700K / i5 6600K (TDP 91W / Idle 2W)

Tcase (CPU temp) = 63C
Tjunction (Core temp) = 68C
Tj Max (Throttle temp) = 100C

5th Generation 14 nanometer: i7 5775C / i5 5675C (TDP 65W / Idle 2W)

Tcase (CPU temp) = 71C
Tjunction (Core temp) = 76C
Tj Max (Throttle temp) = 96C

4th Generation 22 nanometer: i7 4790K (TDP 88W / Idle 2W)

Tcase (CPU temp) = 74C
Tjunction (Core temp) = 79C
Tj Max (Throttle temp) = 100C

4th Generation 22 nanometer: i5 4690K (TDP 88W / Idle 2W)
4th Generation 22 nanometer: i7 4770K / i5 4670K (TDP 84W / Idle 2W)

Tcase (CPU temp) = 72C
Tjunction (Core temp) = 77C
Tj Max (Throttle temp) = 100C

3rd Generation 22 nanometer: i7 3770K / i5 3570K (TDP 77W / Idle 4W)

Tcase (CPU temp) = 67C
Tjunction (Core temp) = 72C
Tj Max (Throttle temp) = 105C

2nd Generation 32 nanometer: i7 2600K / i5 2500K (TDP 95W / Idle 8W)

Tcase (CPU temp) = 72C
Tjunction (Core temp) = 77C
Tj Max (Throttle temp) = 98C

Previous Generation 45 nanometer: i7 860 / i5 750 (TDP 95W / Idle 12W)

Tcase (CPU temp) = 72C
Tjunction (Core temp) = 77C
Tj Max (Throttle temp) = 100C

Previous Generation 45 nanometer: i7 920 D0 (TDP 130W / Idle 12W)

Tcase (CPU temp) = 67C
Tjunction (Core temp) = 72C
Tj Max (Throttle temp) = 100C

-> Core 2

Legacy 45 nanometer: Q9650 E0 (TDP 95W / Idle 16W)

Tcase (CPU temp) = 71C
Tjunction (Core temp) = 76C
Tj Max (Throttle temp) = 100C

Legacy 65 nanometer: Q6600 G0 (TDP 95W / Idle 24W)

Tcase (CPU temp) = 71C
Tjunction (Core temp) = 76C
Tj Max (Throttle temp) = 100C

 

[Sweepr]

Test results for Broadwell and Skylake

The Skylake has increased the number of execution units for integer vector arithmetic from two to three. In general, the Skylake now has multiple execution units for almost all common operations (except memory write and data permutations). This means that an instruction or micro-operation rarely has to wait for a vacant execution unit. A throughput of four instructions per clock cycle is now a realistic goal for CPU-intensive code, unless the software contains long dependency chains. All arithmetic and logic units support vectors of up to 256 bits. The anticipated support for 512-bit vectors with the AVX-512 instruction set has been postponed to 2016 or 2017.

Intel's design has traditionally tried to standardize operation latencies, i. e. the number of clock cycles that a micro-operation takes. Operations with the same latencies were organized under the same execution port in order to avoid a clash when operations that start at different times would finish at the same time and so need the result bus at the same time. The Skylake microarchitecture has been improved to allow operations with several different latencies under the same execution port. There is still some standardization of latencies left, though. All floating point additions, multiplications and FMA operations have a latency of 4 clock cycles on Skylake, while these were 3 and 5 on previous processors.

Store forwarding is one clock cycle faster on Skylake than on previous processors. Store forwarding is the time it takes to read from a memory address immediately after writing to the same address.

Previous Intel processors have different states for code that use the AVX instruction sets allowing 256-bit vectors versus legacy code with 128-bit vectors and no VEX prefixes. The Sandy Bridge, Ivy Bridge, Haswell and Broadwell processors all have these states and a serious penalty of 70 clock cycles for state switching when a piece of code accidentally mixed VEX and non-VEX instructions. This annoying state shift and penalty has been eliminated on the Skylake. Apparently, the implementation of 256-bit registers has become more streamlined.

More here: www.agner.org/optimize/blog/read.php?i=415

Happy new year AnandTech.

 

[VirtualLarry]

I'm still having problems with my G4400's iGPU. This time, the HDMI audio out.

I let the monitor go to sleep, and power off. Then, I come back, turn the monitor on with the remote, then move the mouse to wake up the PC's HDMI output.

Well, it won't sync with the audio portion of the HDMI signal. In fact, going to Playback Devices, in Win7 64-bit SP1, it doesn't even show the HDMI output anymore!

I tried turning the monitor off and on again, no difference.

So then, I went to Device Manager, disabled "Intel Display Audio", and then re-enabled it, and it came back under Playback Devices, then I had to select it and manually click "Set Default".

Then I got my audio back.

Now, this issue isn't new to me. My monitor, a 24" Westinghouse 1080P HDTV, had this issue with my other PCs, when I first got it. Subsequent AMD and NV video drivers, in the following month or two, seemingly fixed the issue, and I don't have problems on my other rigs any more.

But it seems, Intel didn't get the memo, or the secret fix or whatever was necessary to get it to audio handshake.

Edit: This link was posted in another thread. Thought I would post it here, and highlight a quote.
http://www.asrock.com/news/index.asp?id=3132

While ASRock SKY OC breathes life into Intel® non K series CPUs, currently it is still exclusive to motherboards with Intel's Z170 chipset. But have faith in ASRock's skillful engineers, sooner or later more Christmas gifts from ASRock are going to be delivered.

Could they be working on "SKY OC", for non-Z170 boards? I guess the SuperMicro board where this all started, was an H170, so that should be possible. What would be really killer, would be if the H110 / B150 boards could also do that. Bazinga! Take that Intel!

Edit: Here's a video, with some SKL BCLK OC benchmarks, of the G4400, i3-6100, and i5-6400.
https://www.youtube.com/watch?v=TqWNyhjhE6I&feature=youtu.be
Predictably, in games, even OCed to 4.62Ghz, the G4400 is slower than a stock i3, in terms of minimum frames. Just can't quite get out of the dual-core / dual-threaded Ghetto.

 

[Burpo]

Not happy with your new rig? What a shocker..

..maybe we'd like to see you get something that will actually work for you without reservation. Those 4 threads come in handy for a lot of things..

 

[VirtualLarry]

Not happy with your new rig? What a shocker..

Intel drivers have show-stopping bugs on a new platform? What a shocker...

(Yes, I guess I should have purchased an i3, right? I mean, that would just magically make their buggy drivers work, right?)

Edit: Just pointing out that Intel's iGPU drivers are FAR less "polished" and have a lot more issues, than AMD's and NV's. That fact isn't new, but I'm sure that the IST doesn't like it pointed out.

Edit: Just for the heck of it, I'm going to swap in the other G4400.

 

[Burpo]

Just can't quite get out of the dual-core / dual-threaded Ghetto.

Your words..
Should have said won't instead of "can't"..

 

[Arachnotronic]

Intel drivers have show-stopping bugs on a new platform? What a shocker...

(Yes, I guess I should have purchased an i3, right? I mean, that would just magically make their buggy drivers work, right?)

Edit: Just pointing out that Intel's iGPU drivers are FAR less "polished" and have a lot more issues, than AMD's and NV's. That fact isn't new, but I'm sure that the IST doesn't like it pointed out.

Edit: Just for the heck of it, I'm going to swap in the other G4400.

You should buy a nice discrete GPU.

 

[VirtualLarry]

You should buy a nice discrete GPU.

I've got a slew of them. I've just been trying to test Intel's new iGPU (HD 510) in SKL, since that's really the "new" feature of these chips.

Anyways. I swapped in my other G4400, and surprisingly, the newer Intel 4352 drivers installed and are running, without BSODing so far. Steam auto-starts, and did an update at some point, so possibly that was triggering the BSOD? It wouldn't happen immediately at login, it was delayed by a few seconds, like it was triggered by some program starting up.

I still have the image corruption on Newegg, though, in Waterfox. Horizonal black lines on white backgrounds in images that shouldn't be there. Even with the new 4352 drivers, that are supposed to "fix image corruption".

 

[ShintaiDK]

Tried another browser than your waterfox?

My i3 6100U doesn't have any issues in Edge or IE11 on Newegg.

 

[VirtualLarry]

Your words..
Should have said won't instead of "can't"..

I wasn't commenting about my rig... I was commenting about the benchmarks in the YouTube video, and how that even with Skylake and "SKY OC", how Intel's dual-core chips are still so limited, gaming-wise, that they can't even exceed the performance of a stock i3 CPU, no matter how high you clock them (on air). Which is exactly what the story was with the G3258.

I didn't get my G4400 CPUs for gaming, specifically, more for desktop usage. So far, I'm not disappointed in the performance, although the bugginess of the new platform is slightly annoying.

 

[VirtualLarry]

Tried another browser than your waterfox?

My i3 6100U doesn't have any issues in Edge or IE11 on Newegg.

It's weird, I've never gotten that type of image corruption in Waterfox on any of my other systems, so it seems specific to SKL, but maybe it's some interaction between Waterfox's hardware acceleration and Intel's fledgling SKL iGPU drivers.

I'm going to try swapping different RAM in first.

Edit: Ok, it's not the RAM, I swapped in a 2x4GB kit of Avexir DDR4-2400, in place of the 4GB Kingston DDR4-2133 stick. Still getting the image corruption on Newegg.

Not getting BSODs though.

Edit: I just tried Firefox 43.0.3 64-bit, still getting the same corruption.
Edit: Tried IE11. Couldn't get any corruption to appear, but by the same token, it was not anti-aliasing my scaled images properly, they had tons of "jaggies".
Edit: Tried Firefox 43.0.3 64-bit, with HW Acceleration disabled. Still has same type of image corruption. Still seemed to be anti-aliasing images, even with HW Acceleration disabled. So, perhaps it's using some DirectDraw API that caused the GPU drivers to scale images.

 

[ShintaiDK]

Dont complain about drivers as long as you use Windows 7.

 

[VirtualLarry]

Dont complain about drivers as long as you use Windows 7.

Why? Windows 7 is explicitly supported.

https://downloadcenter.intel.com/do...cs-Driver-for-Windows-7-8-1-10-15-40-6th-Gen-

SUPPORTED PRODUCTS:

SOFTWARE

On 6th Generation Intel Core and Intel Mobile Xeon processors and related Pentium/Celeron:

Microsoft Windows* 10 64-bit

Microsoft Windows* 8.1 64-bit

Microsoft Windows* 7 64-bit, 32-bit

*Other names and brands may be claimed as the property of others.

 

[ShintaiDK]

The entire platform got minimal Windows 7 support. Like it or not.

You got a free upgrade to Windows 10. So use it.

Even Microsoft doesn't support Windows 7 anymore outside security fixes.

And did you test another browser?

 

[Maxima1]

The entire platform got minimal Windows 7 support. Like it or not.

You got a free upgrade to Windows 10. So use it.

Even Microsoft doesn't support Windows 7 anymore outside security fixes.

And did you test another browser?

Not to mention, Windows 10 is superior. Heck, I don't get the gripe with 8 either.

 

[VirtualLarry]

Edit: I just tried Firefox 43.0.3 64-bit, still getting the same corruption.
Edit: Tried IE11. Couldn't get any corruption to appear, but by the same token, it was not anti-aliasing my scaled images properly, they had tons of "jaggies".
Edit: Tried Firefox 43.0.3 64-bit, with HW Acceleration disabled. Still has same type of image corruption. Still seemed to be anti-aliasing images, even with HW Acceleration disabled. So, perhaps it's using some DirectDraw API that caused the GPU drivers to scale images.

I'm also getting image corruption, on the background tile images on this forum, it's not limited to Newegg.

 

[ShintaiDK]

So IE11 works, Waterfox doesn't.

Try Chrome as well if you want to be sure.

 

[VirtualLarry]

So IE11 works, Waterfox doesn't.

Try Chrome as well if you want to be sure.

Well, I don't use Chrome. But I'd hardly call IE11 "working", since it doesn't anti-alias my scaled images, and they look horrible. I have a feeling, though, that there's a driver bug in that API routine, though. Since Waterfox / Firefox use that API, but IE11 doesn't, judging solely by the presence or absence of anti-aliased images.

Edit: Update, I tried a bootable USB stick with Linux Mint 17.3 Cinnamon 64-bit on it. Booted very quickly! No corruption in Firefox in Linux. Of course, Linux doesn't use Windows' drivers.

So then I rebooted back into Windows, and I noticed that the scrolling was tearing and shimmering a lot. Although, the images were still anti-aliased. Apparently, you have to re-start the browser for HW Acceleration to fully enable or disable. So my prior test was somewhat invalid.

With HW Accelleration truly disabled, I didn't notice any more background image corruption. Success? But now scrolling is tearing and isn't smooth. Sigh.

I did swap the original CPU back in, and now I'm not getting BSODs. I also have the Avexir 2x4GB DDR4-2400 (@2133) instead of the Kingston DDR4-2133.

I have half a mind to swap the old RAM back in, and see if the BSODs come back.

Edit: I took one of the 4GB Avexir sticks out, and left one in, and it was fine. So then I removed that stick, and put the Kingston 4GB DDR4-2133 back in. Was able to boot, then I used it, seemed fine, then I rebooted. Steam updated, and then BOOM! BSOD again! So maybe something to do with Steam? I took the Kingston out again, and have one of the 4GB Avexir sticks in. Seems OK again?

Edit: Oh, before I forget. I had internet radio playing, and I went into the other room, and came back, turned the monitor back on, and moved the mouse, and trying hitting SHIFT on the keyboard, still no monitor signal wakeup. Pretty-much the same problem that plagued AMD drivers, especially the 7790 / R7 260X / Bonaire. First time it happened, I couldn't even shut it down by hitting the power button. Second time, it powered off OK. It wasn't hung hard, because the HDD light kept blinking, due to it continuing to crunch BOINC in the background.

So, Intel drivers need to improve on SKL. I look forward to a new driver in a week or two.

Edit: It seems like not waking up from monitor sleep mode is a thing with the 4352 drivers. So I've reverted to the 4300 drivers.

 

[Magic Carpet]

Larry, why don't you try Windows 8.1 or Windows 10 instead and see if it fixes at least some of your issues. Don't you have a spare SSD to quickly try it out? If you don't succeed at first, try again. Thought you liked tinkering with your gear.

Not happy with your new rig? What a shocker..

A sign worthy material. Heh.

 

[escrow4]

Always match the vintage of the OS to the vintage of the hardware. Win 7 is 2009/2010. Use 2010-2012 hardware if you want smooth sailing. Skylake is this year. What else was released this year? Win 10. Its like using Win 95 with a 5960X. Not a good idea.