djGrrr

Congratulations on 1 million subscribers

The aluminium build.

Linus, I guess what I'm saying is that you should be in the apology video as well, explaining your part (and not just have it part of the video description). I do not believe Slick to be totally innocent in this whole situation, but all he did was mistake which switches were being used on the specific keyboard he was looking at, he made no claim they were fake.

Yes and in the original video this is exactly what Slick said, he made no claim that any keyboards they sell have fake switches, just different switches, he in fact said that these alternate switches are NOT fake or counterfeit in any way, so him apologizing for that seems silly to me.

I in no way said anything that could be construed as an attack on Armageddon, I didn't even mention them for god sakes. All I said was actually more against you than anyone else. I think its unfair that the video implies that this was all slicks fault, when in fact it was your title that made the suggestion of anything being fake or counterfeit, not slick at all.

Wow, my post defending slick was deleted? So now LTT is stooping to censorship? Well, you've lost my respect.

You do know that you can price match things at NCIX?

Well to be clear, external audio does not need any PLX chips anyhow, as the x1 slots on the motherboards come from the chipset, and not the cpu :) the plx chips are only needed if you are going to want to connect more than 2 high bandwidth (ie x8 or higher) cards into the motherboard

It's not "only good" for multi-way CrossfireX/SLI. Its also very good if you have additional add-in cards. Like if you have Dual SLI/Crossfire and a x8 RAID card, the PLX chips allow you to actually use the 16 lanes from the CPU effectively. No video card is going to come even close to using the bandwidth of x8 3.0, so if you have 2 cards in SLI/Crossfire there is still a tonne of bandwidth left over that can't be used unless you have these PLX chips.

Well its not like they could have changed it to using QPI at the last minute or something, it takes more than a year to change things like that. I am leaning more towards the intel spec pages being wrong. Everywhere except the intel website where i see the the system bus mentioned, it is a DMI bus, not QPI: http://www.hardwarezone.com.sg/product-msi-z87-gd65-gaming http://www.overclockersclub.com/reviews/intel_core_i7_4770k/ I really don't think that everyone except intel has it wrong, i think it's the other way around

Then why is it that all the specs say it's using DMI? http://www.pcper.com/reviews/Processors/Haswell-Review-Intel-Core-i7-4770K-Performance-and-Architecture/Z87-Chipset-Chang https://en.wikipedia.org/wiki/Haswell_%28microarchitecture%29

Does anyone else see the fatal flaw in the new Z87 chipsets? The interconnect between the CPU and chipset is unchanged at a DMI 2.0 bus limited to 20Gbps. Yet we magically have 1.5x the bandwidth simply for the 6 SATA3 ports (36Gbps). This is not even taking into consideration all the other IO on the chipset, such as 6xUSB 3.0 ports (30Gbps), 14xUSB 2.0 ports (6.72Gbps), the 8x PCIe 2.0 lanes (32Gbps), and the gigabit network (1Gbps). So that is a total potential of around 105Gbps of bandwidth, all being thrown through a 20Gbps interconnect. Now granted, using all this potential at once is never going to be feasable anyways but, even in a realistic scenario, such as having 4 (or even 6) SSDs in RAID 0, and maybe even a USB 3.0 device or 2 connected, is going to be severly bottlenecked by the 20Gbps DMI 2.0 bus connecting the CPU and chipset. So i think anyone expecting to actually make use of those 6 SATA3 ports is going to be pretty disappointed with their performance.

My Twitter: https://twitter.com/djGrrr

Would be awesome to win ^^

No, pretty much any x1 slots on a Z77 board are pcie 2.0, and come from the chipset, not the CPU, so the bandwidth is not shared with the gpu at all