*** Official Clevo W230SS/Sager NP7338 Owner's Lounge ***

Follow up Linux guy here.

I do have a single dead pixel--not worth it for me to do anything about it. It's off-center, too, so I had to search to even find it.

My battery results with Linux so far--I've only been getting 3-4 hours. At this point, I don't know why--my w110er got the same battery life, and actually significantly improved under Linux. I will try and see if it is conditional to my machine, but I'm figuring that there must be an issue with the sleep-state in the GTX860m, seeing as it is new and not properly handled in Bumblebee just yet.

Here is a graph of my discharge rate.


As you can see, taking notes in class, wifi on, bluetooth on, and screen at 15%, I was averaging a bit above what I expected. I am going to probably do more testing with bluetooth inactive.

 

If there was a way to get the W230SS to last 6-7 hours without running the battery below 20%, I would buy it.

I intend on undervolting and underclocking the i7-4810mq to 1.2 GHz and disable two of the cores when I'm using it for college, and then set it back to default during the breaks (such as Thanksgiving or the winter/spring/summer break).

This is coming from someone who used a laptop with a Sandybridge i3 clocked at 1.3 GHz. And if it enters the battery mode, the CPU is locked at 800 MHz regardless of the CPU load due to Samsung's locked down BIOS. And I use the battery mode quite often.

The only saving grace was that I installed an SSD in the laptop, though the CPU often bottlenecks the SSD.

The most demanding things I ran was Autodesk Inventor on VMWare (remote desktop connection), on the 800 MHz i3 dual core.

Such power.


(EDIT: I also use Waterfox. The version 28 supports AVX2 extension, which is only available on Haswell, Broadwell and Excavator as of now)

 

Htwingnut was getting almost 6 hours out of the w230ss without any drastic measures on a light browsing/workload test. So you shouldn't have much problem extending that.

 

I wouldn't count on it. About 5.5 hours of actual web use is realistic to about 10%. You really can't get much better than that. It runs at about 9-10W drain on battery with light surfing, about 8W at idle at near lowest brightness, which considering the hardware is pretty good. The Razer Blade 14 gets about 6-6.5 hours, but it has a 72WHr battery compared with the 61WHr in the W230SS.

7 hours to 20% would require either an 80WHr battery or somehow miraculously get down to a 6-7W drain which my i5-4200U Sony netbook barely achieves.

 

Could I ask what you are using to measure your power drain, HT? EDIT: It's just that with 61.2W/hr at 8 watts draw you get 7.65 hours of use, at 9 you get 6.8, at 10 6.12, at 12 5.1hrs at 13 4.7 and at 15 4.1 hours.

Now, in general my software has been reading a discharge rate of between 12-16 at very light loads, wifi on, brightness low, 2.5" Intel 335 and mSATA Crucial 500 with the 4810mq. Do you have any input on that? Thus far I've been getting south of 4 hours in Windows and Linux.

 

The actual wattage makes no difference, it's the times and no tweaks are going to adjust that much.

 

I'm not sure what you're trying to say here. Wattage is everything when measuring battery life.

 

I did an actual test using Firefox and reload every to reload a handful of websites. every 1-3 minutes with browser cache turned off.

Also, are you running power saver mode?

Your calculations are right, but you also have to realize taking battery down to 10% is what you have to consider, so take 10% off those amounts and that's what you get. 6.12 hours - 10% ~ 5.5 hrs.

Actual run time is the best measurement. I actually did an automated light browsing test to get those results. That was the time it took for it to go from 100% to 5%.

I do use BatteryBar though to get a gage on power consumption.

 

Hm, Notebookcheck stated that the Razer Blade 14 gets a little less than 4 hours with the WiFi enabled and pages refreshed every 40 seconds. Idling without WiFi and minimal brightness only gets just above 5 hours.

http://www.notebookcheck.net/Review-Razer-Blade-14-Notebook.107065.0.html

 

Part of the problem is the number of factors that will change battery life.

 

I've had a small amount of experience in having to work inside a power budget run time test is only how you validate your assumptions. Counting watts is how you get there. You have to find the biggest power hogs and figure out if you can use something else or how to minimise the power usage at any given time. The biggest hog of all is the display. Always on always sucking juice. A spinning rust drive eats your battery at a pretty constant rate too. The CPU uses surprisingly little power unless you, or the os or anti-virus, is hammering the system.

The plan to under volt and heavily under clock the CPU should make a difference in this regard.

Can't do anything about the display but turn down the brightness. The fact that it's a smallish panel will work in your favour in that regard. Get rid of a spinning rust drive and stick an energy sipping ssd in it's place (I believe the Samsung evo is one of the most efficient msata drives out there). Keep as much in ram as you can. Most of all limit processes that hit the system. Anti virus or search caching. These will literally suck the life out of your battery

 

I plan on moving the M500 240GB SSD from the i3 laptop to a new one, install a HDD in the i3 laptop, and then donate it to someone.

Not exactly sure how much would the M500 SSD extend the battery life compared to the standard HDD. 20 minutes? 30?

 

Probably quite a bit:

for instance a WD 750GB Scorpio Black 2.5" drive uses a peak of 1.15A@12VDC ~= 6W and at idle ~0.8W (source WD data sheet).
a 250GB Samsung EVO 840 mSATA uses 2.17W Max and at idle it uses 0.21W (average) (source tomshardware.com)

that's a huge delta for a laptop on a battery, probably 30 min increased battery life just from having no spinning rust drive.

 

With my last laptop, an HP dv6t 7000, I was only able to get between 3 to 4 hours max out of the oversized 9 cell battery when I had the original HDD installed. I switched to a 960gb Crucial M500 and my run time jumped to between 4 hours 30 min to 5 hours 45 min depending on power settings and load. It definitely made an immediate and noticeable increase in battery life. I started with a Samsung Evo mSATA right from the beginning with my W230SS so I can't tell you what the difference would be on this laptop, but I'm guessing it would be would have similar results.

 

I also noticed that on Perma's website, someone mentioned that their W230ST's idle power consumption dropped from 15W to 10.8W with v1.1 BIOS, and then down to 8W with v2 BIOS.

If the same power optimization can also be done on the W230SS with future Perma's BIOS mods, I might go ahead and buy the laptop.

 

Yep and here's food for thought. The Maxwell 860m in the new spin is more powerful and more power efficient than the previous model. More control over the fan is a good thing too. It's going to be a long 3 weeks.

 

I assume this means that the ASPM feature has made it out of Beta and is fully stable :thumbsup:.

At stock, the Maxwell chip theoretically has much finer grained power gating coupled with lower level transistor optimizations than the Kepler GTX 765, which I can believe accounts for a good portion of the 6W difference (16W idle W230ST vs 10W idle W230SS) reported by Notebookcheck. With Prema's ASPM feature, both machines should reach parity at 8W idle since this shuts off the GPU when not in use. I.e. the GPU no longer becomes a factor in the idle power budget. The remaining 8W, I assume is divided amongst the screen (about ~3W), the CPU (probs around 2-3W), Storage (~.5W if SSD) and the various onboard ICs + RAM (~1W though I'm aware Haswell can totally power gate the RAM bus)

I was one of the guinea pigs for this ASPM feature on my W110er and idle power dropped to about 9W (Ivy was less efficient at idle).

 

I won't be using the dedicated GPU for school work, which is also why I planned on severely underclocking/volting the CPU and shut off half of it.

Is there also an option to adjust Haswell's IGP's clock rate and voltage?

 

That's already internally managed by the cpu at idle, only load clocks are changeable.

 

The W230ST & W230SS Prema Mods have all ASPM options available, people just need to push a few buttons...

 

Yes display is a big factor, but once you drop below 60% brightness there's less than a Watt difference between 60% and lowest (0%). Undervolting and underclocking your CPU is good and all, but it already runs at 800MHz on Power Saver. You can calculate until you're blue in the face but it doesn't mean squat until you do the actual test. And most users want a realistic and simple scenario where they can just set their computer to power saver mode, screen to 30% brightness, and run with it. Optimizing it is great, sure, but most people won't fuss.

Many HDD's actually consume less power than many SSD's. You need to look carefully in this regard. But in any case, when you unplug you can set your drives to shut off after X amount of minutes, so it's kind of a moot point unless you plan on doing a lot of disk activity on battery.

Even with Wi-Fi off, and screen at minimum brightness, the best I've been able to get is still over 8W draw. I doubt you'll be able to squeeze much more out of it than the already 6 hours at light use.

 

Well darn it. I guess I should buy an extra battery, or keep looking.

 

The HDD in my old laptop must have been majorly inefficient. I have seen some HDD specs that seem to have a pretty low power consumption under load compared to others, but my only experience is with whatever the default HDD was compared to the M500...which definitely made a big difference in my case.

 

Do you have any source for this? I've never known this to be the case and with any mechanical drives, plus if you're using spin-down then you're going to have to wait for it to keep spinning back up (which in itself will cause a power spike).

 

Find me another laptop that is a 13incher with a Maxwell gtx 860m and 1080p IPS display that is now capable of 5-6hrs battery? EXACTLY

 

SSDs do vary in power a lot, worth checking out before buying.

 

I've just checked some reviews and the figures are a bit insane. For example Samsung claims the 840 series uses 0.74W active, yet benchmarking shows 2.5W or so. Not exactly sure how they justify that.

A valid counterpoint for SSDs vs Mechanical though was that even if on occasions SSDs use more power, it's only when reading/writing which they would complete considerably quicker than a mechanical drive so in the long run it would still work out less power. Additionally they can go into deep sleep and back out again without a several second long spinup delay.

 

Tech Report and AnandTech have significantly different reported values from their review on the SSDs.

Crucial's M500 SSD reviewed - The Tech Report - Page 9 (Samsung 840: 1W idle, Crucial M500: 1.1W idle)

AnandTech | The Crucial/Micron M500 Review (960GB, 480GB, 240GB, 120GB) (Samsung 840: 0.3W idle, Crucial M500: 1W - 1.1W idle)

 

Anyone know when the sale on the i7 4810mq ends? Might not be able to order until next month... really don't want to miss that deal though

 

As far as I know, Intel 530 ssd is the most power efficient among the others. It uses 20nm technology.

 

Its not really such a great deal- for example my stock i7 4810qm reached only 3.1 ghz while 4 cores were under load[ and scored about 6,70 in CineBench 11.5)- It's worse result then 4700mq . I managed to raise clocks to 3.6 GHZ which is a proper clock for 4810 according to Intel Specification, but its a bit too much for cooling and the cpu throttles in games like bf4 or coh 2 by 1-3%.

In order to reach 3.6 or even 3.4 which is a reliable frequency you have to increase Processor Current Limi from 60 to 80-85A in Intel Extreme Tuning Utility. I dont know why there was such a limit- 60A, when it is definietely not enough for this CPU...

 

Did nothing, all stock.


I am unable to find "Processor Current Limit" in XTU!

 

I'm looking at picking up a W230SS or W350SS for gaming from my GF's house and at work, mainly for WoW and Diablo 3. I'm looking to run a SSD with the i7-4810.

Are there any heating issues with the 230 I should be concerned with? The laptop may be plugged into a TV when gaming at her house and I'm not sure if I should go with the 13.3 or 15.6 model. My last laptop was an M11x that this is replacing and I absolutely loved the size of it.

 

First of All - Star Stress Test for CPU in XTU then print screen CPU Frequency - Your pictures were made while in idle mode.
I updated my bios with prema mod- maybe thats why I can adjust Processor Current Limit in my XTU...
Also- download cinebench and perform benchmark for 64bit CPu.

 

Heya guys, I got my W20SS / MYTHLOGIC Chaos 1313-A so I figured I would drop in and partake in the forum.

CPU: i7-4810MQ
GPU: GTX 860m
RAM: 16GB Corsair Vengeance 1866Mhz
SSD: Samsung 840 PRO 256GB
WIFI: Killer Wireless-N 1202
OS: Windows 7 Home Premium

First off, love it love it love it! This is my first anything with a SSD so I am quite impressed. My old rig was a ASUS G72GX 17in 9800 Intel dual core, 6Gb, GTX 260m. I have heard of cooling mods such as the "Foil Tape mod" between the heatsink and the fan, and adding copper heatsinks to the piping from the CPU / GPU to the main heatsink, I have not encountered any heat related issues (It lives on a cooling pad, and when I game I turn the on-board fan to max) but I am all about performance and tinkering. Have you guys seen benefits from these modifications? And I hear numerous things about "Prema BIOS", I know Mythlogic PC does some tuning and BIOS stuff of their own, would I be better off with the PREMA stuff? Noticeable performance increase? Lots of questions I know, which must be annoying from a new member, but one of the main reasons I went with this model (other then wanting something COMPLETELY different from my old rig) was the fan-base seems to LOVE these little monsters (I have been haunting anything from HTwingnut for awhile now). Happy to be here and look forward to future conversations!

P.S. What the **** is Arma III's problem?! I toss anything else in my Steam folder at this beast and it devours it, guess I'll be sticking with Arma II for awhile more lol

 

Sager has not provided any end date which is pretty typical of their sales. One day they will tell us its over and we'll have to remove it.

 

Oh I agree that optimisation is more of a pain in the rear than most (including me) want to tackle, but in the case I was responding to the user expressed an interest in shutting off half the CPU and limiting the CPU to 800mHz regardless of load, in such a case a Watt here and there makes a huge difference.

Regarding your HDD v SSD comment revise many to some and I'll agree whole-heartedly with you (mostly because there are a handful of SSDs that have woeful power management regimens). I also remember a certain company that had green drives whose power saving routines were so aggressive that the drive literally wore itself out within 12 months just from spinning up and down the platters. You can't really predict when a major, non RT and non embedded, OS wants to hit the drives though so the lower peak Wattage the better.

62W/h, is that total capacity or capacity with 20% internal reserve? i.e. is that the point at which the lithium turns to snot and the battery is dead? Or does this number indicate a safe power drain from the pack? the battery version of a hard drive gigabyte?

Anywho it'll be interesting if anyone does go after the long-runtime crown

 

@Cadder I get 3.19 ghz when all 4 cores are loaded. Not sure how you think a 4810mq is worse than a 4700mq.
most 3d games and applications are gpu limited not cpu anyways

Review Clevo W230SS (Schenker XMG P304) Barebones Notebook - NotebookCheck.net Reviews
Scroll to CPU performance, 14% increase to 4700mq in single thread. Slight increase in multi-thread. They used cinebench too

 

I get 3.09 for quad core load... 4810qm is a much better processor then 4700 on paper and it should run @ 3.6 - just look at multipliers in XTU 38 for single core, 37 for dual core and 36 for 3 and 4 core then why is it runing at 3.2 / 3.1 . Whats the point of installing 4810qm when it has same clocks as 4700?

 

Someone should do an OCing and undervolting study between a sample of the 4700s and 4810qms.

The only reason I'm considering the 4810 if I buy a W230ss is because it's probably better binned. Aka, less voltage for the same clock rate.

 

Arma is notorious for its unoptimised code.

 

And the sale that it's on in NA ($35 for the better processor) isn't half bad either.

 

You'd need a batch of each to get a subjective evaluation. Otherwise its just conjecture. Most people won't have the budget needed for that many chips .

Sent from my DROID RAZR using Tapatalk

 

4712 was the best choice for me, hopefully it will get a little less hot since it's 37W (like a core i5).

If someone with a 4810 could give me the same test they did, we could compare the temperatures. See if it really changes something. Won't be the same conditions, but still better than nothing.

My w230ss comes tomorrow, can't wait

 

Maybe several of us that have 4810s and the 4700s can do the experiment, starting off with finding the minimal voltage needed to maintain a 12-hour stable 2.4 GHz.

I'll do it if I buy the W230SS. Though I won't make that decision until after the college finals week (next week) is over.

 

The Clevo download page for the W230SS only lists Windows 8.1 drivers ( CLEVO - Download) nothing for Windows 7. Do those same drivers work with Windows 7 or are they located elsewhere? (or do you have to source your own?).

 

My BIOS Mod not only effects the BIOS but also adds a few thing things to XTU & Throttlestop...

Mythlogic is a ' Prema Mod Partner', when purchasing from them you already get a special version of the Mod BIOS customized for Mythlogic's needs with all the bells & whistles and covered by warranty.

 

I have the XMG 304 with 4702MQ processor, which has been really great so far apart from the occasional loud fan noise. Is there anything I can do about it? Maybe with prema mod? It seems to be almost off or kick in very loud when ever I have too many programs open. I don't even mind about lowering the performance a bit as long as the fan noise would stay a bit more quiet or stable.

 

I'm thinking that perhaps you could cut the fan cable and stick a resistor in-line. Basically like a home-made mod of what the Noctua fans normally come with, as I think we have a bit of thermal headroom. I'm not sure what resistor would be suitable though given what voltage we have to play with.

 

Looks like a new bios has arrived for the schenker xmg p304 which I believe is a rebranded w230ss