The Haswell Ultrabook Review: Core i7-4500U Tested
by Anand Lal Shimpi on June 9, 2013 9:00 AM ESTBattery Life
I didn’t get my hands on the Haswell S7 until a few hours after I landed in Taipei. My hotel’s in-room internet was throttled to around 2.5Mbps, which wasn’t enough bandwidth to reliably run our web browsing battery life tests. Luckily, I had my review copy of PCMark 8 on hand with its new built in battery life tests. I asked Jarred to run comparison data on the Ivy Bridge S7.
We had time to perform multiple runs on two of the tests: Home and Creative.
From the PCMark 8 Technical Guide:
The PCMark 8 Home benchmark test includes a set of workloads that reflect common tasks and activities performed by a typical person at home. These workloads generally have low computational requirements making the PCMark 8 Home benchmark suitable for testing the performance of low-cost tablets, notebooks and desktops.
The PCMark 8 Home benchmark test contains the following workloads: Web Browsing, Writing, Casual Gaming, Photo Editing and Video Chat.
The PCMark 8 Creative benchmark test includes a set of workloads that reflect tasks and activities typical of more advanced home computer users. With more demanding requirements than the Home benchmark, the PCMark 8 Creative test is suitable for testing the performance of mid-range computer systems. Your system must have a GPU with full DirectX 11 support in order to run all the workloads in the PCMark 8 Creative benchmark.
The PCMark 8 Creative benchmark test contains the following workloads: Web Browsing, Photo Editing, Batch Photo Editing, Video Editing, Media to Go, Mainstream Gaming and Group Video Chat.
As always, I calibrated both displays to the same brightness (200 nits). In the case of the Haswell based S7, I disabled all additional display power saving options in the Intel driver. Keeping in mind the new S7 has a 33% larger battery, I’m presenting both absolute battery life numbers as well as minutes per Wh for normalized comparisons.
The PCMark 8 Home battery life test is the lighter of the two, and thus has the best chance of showing peak improvement on Haswell. The results are very good:
| PCMark 8 Battery Life | ||||
| PCMark 8 Home | PCMark 8 Home (Normalized) | PCMark 8 Creative | PCMark 8 Creative (Normalized) | |
| Acer Aspire S7-391 (Core i7-3517U) | 2.83 hours | 4.857 mins/Wh | 3.35 hours | 5.743 mins/Wh |
| Acer Aspire S7-392 (Core i7-4500U) | 5.2 hours | 6.783 mins/Wh | 5.12 hours | 6.674 mins/Wh |
| Haswell Advantage | 39.6% | 16.2% | ||
In both benchmarks, Haswell ULT delivers 11 - 14% better performance and substantially longer battery life. Normalized for battery capacity, Haswell ULT offers 16% better battery life in the Creative test and almost 40% better battery life in the Home test. Note that the performance advantage pretty much disappears once we move to the Balanced power profile with the laptop connected to the wall.
| PCMark 8 Performance | ||||
| PCMark 8 Home (Power Saver) | PCMark 8 Home (Balanced) | PCMark 8 Creative (Power Saver) | PCMark 8 Creative (Balanced) | |
| Acer Aspire S7-391 (Core i7-3517U) | 1595 | 2694 | 1391 | 2508 |
| Acer Aspire S7-392 (Core i7-4500U) | 1777 | 2832 | 1583 | 2553 |
| Haswell Advantage | 11.4% | 5.1% | 13.8% | 1.8% |
Update: I made it back to the US, equipped with decent internet speeds, I was able to run our light 2013 web browsing battery life test. The increase in battery life is tremendous:
The new S7 delivers over 2x the battery life of the old model. Normalizing for battery capacity, the improvement due to Haswell is 57.5%. These results track perfectly with what we saw in PCMark 8. Workloads with greater idle time will show the biggest improvement in battery life thanks to Haswell ULT.
Facebook
Twitter
RSS
87 Comments
View All Comments
Homeles - Monday, June 10, 2013 - link
Maybe you didn't catch on with Anand's "I was running this in my hotel room" statement, but the idea was to get the anxiously awaited battery life numbers out to the public.seapeople - Tuesday, June 11, 2013 - link
You mean a hotel room in Taiwan is not the normal procedure for a pedantically complete review?ciparis - Sunday, June 9, 2013 - link
OT: Opening paragraph typo: "Haswell less than a month after the arrival of a new CEO,"jhoff80 - Sunday, June 9, 2013 - link
I know you said that you disabled any Display Power Savings options in the Intel driver, but still, out of curiosity, it would be interesting to know what kind of effect those have. I mean, it wasn't made explicit, but does this specific ultrabook support Panel Self Refresh? If so, what improvements does that give?yoyoma245 - Sunday, June 9, 2013 - link
I don't understand why battery life increased going from pcmark8 home to pcmark8 creative. Wouldn't a more demanding test suite result in reduced battery life?meacupla - Sunday, June 9, 2013 - link
Could Minecraft be added to benchmarks for ultrabooks?HD5000 is obviously quite pitiful at eye candy games, so how about popular games that are most likely to be played on them?
I get around 34~40fps with surface pro, which is playable, but could be better.
esgreat - Sunday, June 9, 2013 - link
I don't think I've seen benchmarks for HD5000 yet. The i7-4500u uses HD4400 graphics.With 2x the EUs, HD5000 should give quite a performance boost, but not as fast as Iris.
krumme - Sunday, June 9, 2013 - link
Haswell U improves excactly where it was needed; on the battery life. This segment dont need more cpu power or gpu power than ib, they want battery life.This is the luxury product that ultrabooks are made for.
Haswell for the desktop was utterly unimportant, but this is excellent targeted and a very tangible improvement for everyone.
name99 - Sunday, June 9, 2013 - link
"Haswell for the desktop was utterly unimportant"Uhh, well apart from defining a rather different parallel programming model going forward...
krumme - Monday, June 10, 2013 - link
Yes. Thats relevant in perhaps 5 years from now. Perhaps. Its a technology and innovative huge step forward, but hardly of any pratical importange to the consumers today.