Yes, I realize the Ivy i5 is 35W and the Skylake i5 is a 15W. It is the general direction that x86 CPU's are going... slightly better IPC per clock, but much better power consumption (hence a lot of complaining about how CPU's haven't really improved much in the last 5 years). Geekbench certainly has its controversies on x86, but of the results in the above link I parsed, the major outliers are AES and SGEMM (which seems like an OpenCL one, so its more testing the GPU?). All the other ones are within an earshot.
Note I'm not saying an Ivy i5 laptop is as good as a Skylake i5 laptop, I'm noting that even though the processing power in the two is about the same, the additional platform benefits that a Skylake i5 would bring (lower power, better IGP, etc) is just as significant.
Yea good point, IGP is certainly a lot improved over the generations. Ivy wouldn't even hold a candle up to Skylake. Broadwell i5-5xxxU is noticeably behind the i5-6xxxU chips as well.
I'd agree with you from experience.
Users are getting duped to buy a new model just for the sake of it being new. A higher end 2-3 year old CPU will normally perform just as good, and clock better and be far cheaper. Except where new features/graphics are concerned.
I have AMD C50, E300, A4-4300m and Intel i7-3667U, i5-5300U and i5-6200U in laptops. The latter 3 are all Lenovo ThinkPads. They have the approx same perf and user experience, except when the graphics get concerned. The display ratings are also different.
The i7 clocks higher and stays there longer if I restrict the TDPs, which is what first alarmed me since it makes it perform better in the end.
In IXTU, I set most of them to 15W/25W, but the i7 can be clocked much higher than base, and even TDPs for when you need. 2.5GHz/3.2GHz it is sustaining with ease and it sustains 2.9GHz all cores when I benchmark it (under stock TDPs) which makes it perform ahead of the other two.
I don't notice any power/efficiency difference between the 5300U-6200U, absolutely none, which was disappointing. However, there is a ~1-3hr average usage gain from the 3667U which is certainly a boost.
In my experience, unless you really need a new feature, an older i7 ULV is a far better choice.
Sent from HTC 10
(Opinions are own)