Skylake - let's wait and pray for it together

Very nice to have, especially if you spring for 128gb of ram. That's some trickster-caliber shit right there.

 

http://anandtech.com/show/9320/intel-broadwell-review-i7-5775c-i5-5765c



AMD pls



If Intel ever had driver issues that caused studdering compared to amd graphics, then that's no longer the case.



And of course, Intel is completely shutting down the competition when it comes to power consumption (and by extension, heat). This is what happens when your architecture is built to play ball in the sub 5w space.



No overclocking results yet, but I look forward to seeing cutress put 1.4+V into that poor chip since that's totally a reasonable thing to do to a 14nm chip. 5ghz, hooooooooo!

EDIT Intel also released their 47w quads. So yeah, the desktop parts are probably just mobile parts with larger tdps. Intel is crazy like that.

http://anandtech.com/show/9326/intel-launches-five-47w-laptop-broadwell-skus

I want to know why Apple updated their mbps with Haswell right before this announcement. That's odd. It must've been intentional. Apple is all up on Intel's dick, so they weren't blindsided by this announcement, I guarantee you.

 

That's what I'm curious about too.

I was originally convinced that the reason why Apple didn't update the 15" MBP at the same time as the 13" MBP in March was that Apple was waiting for Broadwell. Perhaps Apple was waiting for the M370X before updating the 15" MBP, but given the timeframe it's strange why Apple didn't just wait another month or so for Broadwell.

 

$366 for the i7, that is less than what I expected. Unfortunately no comparisions at same clockspeed from Anandtech, I want to know if Broadwell desktop are stuck at 5% IPC increase like the mobile chips.

 

oh wow double the price for 5fps increase

 

A fella in the Anandtech comments speculated that out was based on price. Similarly, he speculated that the gpu choice had price-goals. Apple is getting old tech at cheap prices.

I think Apple has excellent relationships with both Intel and amd, so I wouldn't be entirely surprised if that was the reason. Apple doesn't get blindsided by stuff.

 

Wow, those chips are already available in Europe. The i7 costs 412€, more than the 5820K...

 

Across all games tested, it's something like a 20% performance bump over the previous #1, but it does it with less power used. And there's kind of an industry-leading cpu duct-taped to that gpu, so there's that.

If Intel had any meaningful competition, they could easily release a cheaper socketed cpu with two cores disabled. They already have a discount mobile chip exactly in that configuration.

Thank God you euros are savers. You'll need to be saving if you want to afford Skylake.

 

I will just wait for some action of islamic freedom fighters on US soil, now that you guys stopped the Patriot Act. That should bring the dollor down.

 

As for the IPC, if we normalize Anandtechs Cinebench scores to 1 Ghz we get:
3.7 Ghz score 157 points --> 42,42 points / Ghz
4.4 Ghz score 181 poins --> 41,14 points / Ghz

If that quick calculation is right, desktop Broadwell has about 3% increase in IPC.

 

If that's our metric, then it's upwards to 4-5% if you compare the cheaper skus, which are closer in clocks, but (probably) still binned similarly.



And naturally, Skylake oughta provide another 5% on that, amirite?



EDIT lelno

 

No, Skylake will provide a 10% increase on top of that. A die shrink generation is something completely different from a generation that gets a micro-architecture optimized for 14nm.

 

Something is fishy with Broadwell, there is still not a single overclocking benchmark out there.

 

Computex just started, so the "part 1"style reviews are to be expected?

Do you think Intel intentionally scheduled their embargo so that the initial reviews (which generally get the majority of the press) would omit overclocking and other time consuming stuff?

 

I think Intel just requested from all reviewers that they do not publish any overclocking benchmarks. Why they did this I don't know, but the most probable guess would be that the chips are poor overclockers.

 

I seriously doubt that they offered a separate embargo for overclocking.

It's really probably just because of computex. There is precedent for this kind of thing. These review sites only have so much capacity abs trade shows eat a lot of it.

Whether or not Intel intentionally exploited that phenomena is another thing entirely. They probably did. But that's just a sly passive way to minimize the presence of overclocking results in "part 1" reviews. There's deniability.

If they outright embargoed overclocking to a later date, then that's a potential pr nightmare waiting to happen. You can't get the benefit of the doubt like the other possibility. There's no way Intel would be that stupid.

 

There is already a cashback deal on the new Broadwell chips - e.g. 50€ off, if you get them with a new ASUS board. Funny thing is, the chips aren't even available yet.

https://toptechcashback.com/at/de/pages/cpumb/qualifying

75€ for 5820K + board
50€ for 4790K + board
40€ for 4690K + board

In my weakest moments I think about just getting a 4690K and be done with it.
Emma, give me the strength to resist!

 

So yeah, that's a pretty fucking terrible overclock. Any IPC gains are swallowed by the poor OC. I hope that's a lemon.

 

what about that chinese dude that tested overclocking in march?
I can't find the article anymore but he was at 4.6 on air.

I wounder if L4 cache makes any difference in benchmarks, when you compare igp vs discrete gfx card.

 

.

It depends what kind of voltage it took to get there. In general, you can't put as much voltage through smaller chips or they asplode.

So if the necessary vcore was much lower than 1.28, then it wouldn't be that bad of an overclock. But that voltage is high-ish even for a Haswell chip at that clock speed, let alone a smaller Broadwell chip.

So yeah, I'm hoping that those guys were either bad at overclocking or got a lemon.