Broadwell ULV performance review and benchmarks

[thedavexp]

CPUz says single channel memory, is this correct?

Correct this version I tested was only 1x8gb where as the one I originally benchmarked was 2x8gb (dual channel)

 

[Arachnotronic]

Thanks again for your efforts, thedavexp. Great stuff.

 

[ikjadoon]

Hey Guys

Sorry about the delay in replying I work retail and its christmas 🙂

I have taken screenshots of all the CPUz windows on another hp laptop running broadwell unfortunately the original laptop I tested has sold out but this one is running the same CPU (without dual channel memory).

https://www.dropbox.com/sh/03tv0lihfrlcw9v/AACEZOTltIbLfxHB2_KzkonPa?dl=0

I can run some more benchmarks in the new year (i will try do a geekbench done on the broadwell laptop we have left before hand)

Thank you hugely for checking. From the serial number, it's DDR3L which is standard fare for laptops: https://www.skhynix.com/products/computing/view.jsp?info.ramKind=20&info.serialNo=HMT41GS6BFR8A I was hoping for broad LPDDR3 adoption, but alas, maybe only higher-end laptops will get it, like with Haswell.

Interestingly, the PCH name says "L-P"; the code-name for the 8-series of chipsets is Lynx Point. But, maybe that's good enough as there was no "Haswell Refresh" for mobile, AFAIK, which use Wildcat Point as their PCH.

I'd love to see if USB 3.1 made it or the display was luckily IGZO, but judging from the DDR3L, it's probably just IPS or maybe TN.

 

[thedavexp]

Just did some testing with geekbench 3 different model to the one tested originally but same cpu
http://browser.primatelabs.com/geekbench3/1600756

 

[thunng8]

Just did some testing with geekbench 3 different model to the one tested originally but same cpu
http://browser.primatelabs.com/geekbench3/1600756

That is a really weak score .. more than 10% lower than Haswell. I at least expected 5-10% improvement, not a decrease.

For example, the i7-4510U scores:

http://browser.primatelabs.com/geekbench3/932469

 

[mikk]

That is a really weak score .. more than 10% lower than Haswell. I at least expected 5-10% improvement, not a decrease.

For example, the i7-4510U scores:

http://browser.primatelabs.com/geekbench3/932469

These scores differ considerably, you picked up an exceptional good i7-4510U result. On the same time his i7-5500U result is the weakest so far--> http://browser.primatelabs.com/geekbench3/search?q=+i7-5500U

i7-4510U has a 100 Mhz stronger singlethread Turbo, if there is no IPC advantage in Geekbench for Broadwell, a slight decrease is logical for the ST Score.

 

[thedavexp]

Yeah the pavilion was a tad slower we got the envy back in so I ran the benchmark again and got a better result: http://browser.primatelabs.com/geekbench3/1607755

 

[thedavexp]

Ran geekbench on a broadwell i5 5200u for comparison: http://browser.primatelabs.com/geekbench3/1607759

 

[mikk]

Yeah the pavilion was a tad slower we got the envy back in so I ran the benchmark again and got a better result: http://browser.primatelabs.com/geekbench3/1607755

Looks much better. Maybe you yould do a 3dmark11 run, would be interesting to compare with Haswell ULT.

 

[thedavexp]

Looks much better. Maybe you yould do a 3dmark11 run, would be interesting to compare with Haswell ULT.

I gave it a go when I had the machine sadly it didn't run I will see if I can get it working over the next few days

 

[greatnoob]

Ran geekbench on a broadwell i5 5200u for comparison: http://browser.primatelabs.com/geekbench3/1607759

That is around the same performance of an i5 2410m @ 2.3ghz (turbos to 2.9ghz single and 2.7 two core):
http://browser.primatelabs.com/geekbench3/70081

(also the highest multicore score for the 2410m is 5300, the above result is where the average lies)

 

[thedavexp]

I am abit sad that another website seems to be taking credit for my work 🙁
http://wccftech.com/broadwellu-inte...utm_campaign=Feed:+Wccftechcom+(WCCFtech.com)

 

[bullzz]

@thedavexp - i think they kinda did. wccftech mentioned sweclockers as source. sweclockers mentioned your channel as source
http://www.sweclockers.com/nyhet/19814-intel-broadwell-for-ultrabook-i-prestandatest

next time add a water mark 🙂

 

[thedavexp]

@thedavexp - i think they kinda did. wccftech mentioned sweclockers as source. sweclockers mentioned your channel as source
http://www.sweclockers.com/nyhet/19814-intel-broadwell-for-ultrabook-i-prestandatest

next time add a water mark 🙂

Ah yeah didn't see the sweclockers article thanks for that 🙂

 

[bullzz]

u still have to add a water mark so others dont add it for u (like how sweclockers did)

 

[MisterLilBig]

So a side by side 20% improvement, in a benchmark, with more EU's and architectural changes?
That's horrible!

 

[greatnoob]

So a side by side 20% improvement, in a benchmark, with more EU's and architectural changes?
That's horrible!

I think the performance gains are much greater than that. Keep in mind these ULV chips are using DDR3L RAM which (I'm guessing) hinders performance quite a bit in terms of iGPU performance. On a desktop or the 'normal' non-ULV mobile parts, there should be a much greater iGPU performance gap.

 

[mikk]

So a side by side 20% improvement, in a benchmark, with more EU's and architectural changes?
That's horrible!

20% compared to what? You overlooked something. Furthermore, Haswell already did pretty good in 3dmark. Real gaming performance was the big issue where it suffered compared to AMD often. Real world gaming performance is unknown on a more unconstrained TDP Broadwell.

 

[witeken]

So a side by side 20% improvement, in a benchmark, with more EU's and architectural changes?
That's horrible!

Lol, in this article it's 60% and amazing: http://wccftech.com/broadwellu-intel-hd-5500-igpu-performance-numbers-revealed/

 

[mikk]

http://www.3dmark.com/3dmv/5180008
http://www.3dmark.com/3dmv/5180013

To a comparable Haswell ULT 34% (GPU Score).

 

[SPBHM]

http://www.3dmark.com/3dmv/5180008
http://www.3dmark.com/3dmv/5180013

To a comparable Haswell ULT 34% (GPU Score).

this HD 5500 GPU score is 4300 points (3200 for this HD4400)
this is close to my old 9600GT (4800), pretty good considering the TDP, but still to slow for current games, ivy bridge/haswell ULV also can't keep the clocks high enough while gaming, maybe the new one can keep higher clocks while also loading the CPU cores, or maybe not.

 

[B-Riz]

I'm confused about the harping of graphics performance on an ULV part. The point of these chips is power sipping.

I have never bought an Intel laptop powered by their graphics thinking its graphics performance was better than good enough for web and office type stuff.

If someone really wants decent graphics they will have to get a non-ULV or look for a discrete added, like a GT 820M or beefier.

 

[Qwertilot]

The graphics is meant to be the major area of potential improvement in Broadwell, hence the modest interest. Probably won't really know what they've done with it until Broadwell K etc arrives of course 🙂

 

[IntelUser2000]

So a side by side 20% improvement, in a benchmark, with more EU's and architectural changes?
That's horrible!

That's disappointing. I speculated that earlier CPU-World article saying that GT3e Broadwell-K is 80% faster is likely compared to Haswell GT2(not GT3e), and looks like that may be the case.

 

[mikk]

That's disappointing. I speculated that earlier CPU-World article saying that GT3e Broadwell-K is 80% faster is likely compared to Haswell GT2(not GT3e), and looks like that may be the case.

It's a disappointment that you can't read these results properly. Difference is higher than 20% in Vantage. In 3dmark11 the difference is roughly 50% despite being clocked lower. 3dmark is not that important, Haswell did quite good there. Gaming performance is an unanswered question.