Still no word on i3's... guess the 65W i5 6600 is on my list but I may wait out the 1151 socket release early on so manufacturers can revise their first release motherboards.
Upgrading to haswell is > 2x already.
i7 720qm is around 2.9 points in CB 11.5 Multi. (current U level chips can do better)
Most haswell chips (4700m) score around ~7.
I am too lazy to read.
So is Skylake a worthy upgrade from my 2600k @ 4ghz?
All told, the combined Fab 28, which could be geared for 10nm finFET production, is expected to have a capacity of 30,000 to 50,000 wafers a month, according to SEMI.
Intel was originally supposed to install the high-volume equipment for that fab this summer. But now, Intel won’t move the tools into the fab until the second quarter of 2016, according to sources in the IC equipment industry.
Intel, in fact, has pushed out the volume ramp for its 10nm finFET fab by several quarters, according to sources. Intel was late in terms of ramping up its 14nm finFETs. The 14nm delay threw off Intel’s 10nm schedule, sources said.
Were you seriously expecting 10nm to roll out on scheduleSo... what comes after a delayed 14nm process node?
A delayed 10nm node!
SemiEngineering 10nm Fab Watch
So... what comes after a delayed 14nm process node?
A delayed 10nm node!
SemiEngineering 10nm Fab Watch
So about a year delay like 14nm it seems? Would that mean 10 nm mid 2017 for the consumer?
Not much impressive tbh ~
https://browser.primatelabs.com/geekbench3/compare/2321616?baseline=2745895
A better gauge IMO would be looking at the best scores from geekbench, like the one I posted, but prelim results aren't encouraging.Comparison:
Using the average score for each chip @ Geekbench 3 32-bit (from Notebookcheck).
Once again, it's a single pre-launch score, future scores and the average results for each chip might be better.
A better gauge IMO would be looking at the best scores from geekbench, like the one I posted, but prelim results aren't encouraging.
You forgot the higher base clock for skylake, like I said previously the prelim results aren't good but I'll wait for better scores to see how good Skylake really is ~You're comparing the highest Windows score using 64-bit Geekbench for the 5500u to a single score using 32-bit Geekbench for the 6500u. Even if only look at the highest windows 32-bit score for the 5500u (2921 and 6152) you'll see a substantial increase for Skylake. And the average for the 5500u is quite a bit lower.
I go for the "best vs best" scores but yes these are prelim results so not much can be drawn from them.I find that a bit unfair. I'm pretty sure we will see much better scores from these chips after the formal launch, by the time it gets as many or more design wins as Broadwell/Haswell right now. That's why I tend to pick average results for these comparisons.
You forgot the higher base clock for skylake, like I said previously the prelim results aren't good but I'll wait for better scores to see how good Skylake really is
I go for the "best vs best" scores but yes these are prelim results so not much can be drawn from them.
Then why go through the trouble of pointing out that the geekbench score was 32bit, the 32 vs 64bit scores are hardly 5~10% apartI did? Who cares if a performance increase comes from a clock speed increase or an IPC increase? Isn't it better performance either way? Further, I'm not sure that a higher base clock matters in a short bench like GeekBench. Seems like turbo clocks are more relevant.
Then why go through the trouble of pointing out that the geekbench score was 32bit, the 32 vs 64bit scores are hardly 5~10% apart
That might be the case with with Perf/efficiency numbers but for solitary benchmarks (like cinebench) I don't see why it'd be a problem, we do that with phone benches (like sunspider, octane, kraken et al) all the time. Besides any two products, same make & model, picked randomly off the shelf will not give you the exact same numbers nor will the same components have the exact same electrical characteristics, leakage & all, just like the way it works on desktops.I don't think "best vs best" is an effective way of comparing mobile processors. So much depends on implementation by the manufacturer.
Then let me rephrase that for you, clock for clock it doesn't look good at this point in time. Is that better or do you not care about the ~5% better base clock that Skylake has, just remember thermals & power numbers are still unknown.Because the 32-bit vs 64-bit scores are 5-10% apart? That's a pretty sizable gap and its fairly consistent across a broad range of Intel processors. Therefore, it is reasonable to assume that a Skylake part would also benefit from the 5-10% jump if benched under 64-bit GeekBench. Your benchmarks need to be the same in order to make the comparison valid.
You're talking about measuring with two different rulers, but the second ruler has 10% more space between the hashes than the first. An object hasn't shrunk merely because you got a smaller measurement using the second ruler.
That might be the case with with Perf/efficiency numbers but for solitary benchmarks (like cinebench) I don't see why it'd be a problem, we do that with phone benches (like sunspider, octane, kraken et al) all the time. Besides any two products, same make & model, picked randomly off the shelf will not give you the exact same numbers nor will the same components have the exact same electrical characteristics, leakage & all, just like the way it works on desktops.
Then let me rephrase that for you, clock for clock it doesn't look good at this point in time. Is that better or do you not care about the ~5% better base clock that Skylake has, just remember thermals & power numbers are still unknown.
I think you're saying the same thing just differently.With phones your comparing the phones--not the underlying processors. The same processor may perform differently in different phones (thermal characteristics of the phone). If you wanted to compare the processors, either do it comparing two processors in the same phone, or do it using an average (or median, or whatever) for the processors over a range of phones.
To which I say ~I don't think "best vs best" is an effective way of comparing mobile processors. So much depends on implementation by the manufacturer.
The thermals & base clock is relevant here because it's a notebook part & there's a slight chance that the turbo clocks might not last that long, highly unlikely but still, hence the hint to a performance plateau.I would think users would care about absolute performance, not necessarily per clock performance. And at any rate, judging per clock performance by using base clocks and short benchmarks is clearly flawed.
I agree more information is needed. But more information is not needed to say that the Skylake part has a performance advantage over the highest 32-bit windows score available for the Broadwell part. Maybe it is because of clockspeed. Maybe it is thermals. Maybe it is IPC. Maybe it is a mixture of the three.