*** Official X8100 (Sager 8120) Owners Thread ***

I seriously might buy one of these, if SLI 470Ms come available.

 

catalyst 10.8
I don't know which version i have, @work right now
All i know it's from juli 2010, will check the build number this evening.

edit:

bios version : 1.00.11
kbc/ec firmware : 1.00.07

 

Unfortunately as I discovered quite painfully with the M980NU [and a huge number of people with the Alienware M17xR1], Nvidia driver support is NOT always better, and thier 'behind' closed doors and attitude to open standards can be downright slimy.

Both companies are swings and roundabouts, with both producing good and abysmal drivers in equal measure. Nvidia certainly don't deserve the 'golden' driver reputation they used to have. (Hell I've seen usually staunch Nvidia supporters panning them on driver performance elsewhere over the last year)

 

Sager has listed on their website the same BIOS version, but with a couple of letters at the end: Version 1.00.11oc, added on Aug 25th. It's probably the same thing, but it's worth trying.

 

oc = overclocking??? Would be nice.

 

It's possible, saying nothing else without confirmation however.

 

Hey Alex, how is your BIOS testing going? All I am waiting is a good word on that... Or maybe just buy it, sell the slave and wait for a fix/lower prices/6870

 

When I heard from them last yesterday all was going well, they were just checking it was ok to format my hdds without backing up,to clear up a driver conflict with the old NVIDIA drivers which had been on my hdds from when the drives had been on an install with the 285s.

What more can I say than soon?

 

Ok as an update...she's ready to blow
Unfortunately I've only just been able to respond to the email I was sent (as Im at work) at 4pm so I probably wont see it til Thursday

According to Kobalt after 'exhaustive testing' they've noticed no graphics issues, and it scores 18k 3dmark 06 straight out the box
Should be an LED/USB3.0 model too [my old one was USB2, CCFL...bonus!]

As I've likely missed the cutoff for delivery tomorrow I'm going to see about whether I'll be able to do the drive and collect, if not it'll be Thursday.

No luck with the overclocking BIOS though, as they put it the new one they've recieved is simply too hit and miss.

 

I think that since CF came out everybody has LED/2xUSB3.0 . The specs list at Clevo's site are dated 25th of May...
do you know which version is the BIOS?

 

Well I specifically confirmed those details, because cards/chassis only recently arrived, but theorectically they should work in an older motherboard too. My last few X8100s were all 2.0/CCFL despite the mention of LED floating round for a few months, so its a bit of a bonus and machine should be that little bit nicer/longer term because of it.
After my previous luck I just wanted the confirmation really!

I dont know which BIOS revision I'm on yet, all I know is that the [new] overclocking BIOS Kobalt tried was 'inconsistant', so I'm likely not on it! Probably .11, but I will check myself, do some benching/testing (obviously observing for any 'wierdyness') and feedback once I've got the machine in my hands.

As I'm not off again past tomorrow til next week, Im kinda hoping I'll be able to get hold of it tomorrow, but I'm not expecting it.
*Fingers and toes crossed for a flawless machine!*

 

Hi Guys

Just a quick update regarding the X-Fire setup and OC BIOS:

After our testing on the new X8100 rev chassis that supports single GTX 480M and X-Fire 5870 (with the addition of 2x USB 3.0 ports) we did initially find similar X-Fire issues that have been reported in previous posts on this thread. For example during our initial testing of BF2 thegame was virtually unplayable and had trouble getting past the menu screen (which had corruption issues). We have now successfully tested the .11 OC BIOS and that immediately cleared the BF2 issue, it now runs very smoothly in game which is 100% better of course. The OC options also seems stable with up to x4 multiplier increase through the BIOS. I have quite a few screenshots which I will post later this afternoon but it does seem like the updated .11 BIOS has solved the majority of the X-Fire issues that have been reported on this board.

On the GPU drivers, we are using the official Clevo version that we are shiping on the driver CD. Didn't have much success with 10.8 but seeing as we have made a lot of progress with .11 and the official Clevo drivers I would recommend sticking with this setup for now.

 

Thanks for the great job, Neil!
Have you noticed any latency issues when using CF?
You could run this tool in the background while opening multiple tabs in the browser and scrolling them up&down, switching between apps, etc. The max spikes should stay under 5k (and they are rare). Huge spikes (40k+) will have an impact on video/audio/gaming experience.

 

We have just shipped out the last of our chassis from the first shipment, though Alexrose1UK has just picked one up so he may be able to run the test. If not I'll be able to run it in a couple of days when more come in

In the mean time here are some screenshots for reference:

(All taken with 940XM, 4GB RAM, 80GB Gen2 SSD and X-Fire 5870)

Stock 3DMark06 18,128

OC 940XM x4 3DMark06 20,424

SuperPI 1M idle with x4 OC

SuperPI 1M during test showing CPU speed

 

Awesome !

Where I can find the .11 OC bios ?

 

Software Upgrade

 

I just noticed in your 3dmark06 runs that the one of the CPU cores runs quite hot. Although I did a retention mod on my CPU, the difference between our CPU temps is quite substantial. In 3dmark06 I top out around 75C at stock and I noticed yours was hitting 87. Might want to check that paste job.

 

^^ Could be because of the higher affinity for that core. Couldn't hurt to check the paste as well...

 

Awww...Neil ruined my fun lol

As Neil has stated, I have the machine, unfortunately due to roadworks and traffic I didnt arrive home til quite a bit later than expected, I'm reformatting the machine with a fresh install of Windows and then I'll be throwing drivers at the machine and seeing how it goes. Unfortunately I have work tomorrow so its not going to be til the evening UK time where I'll start testing proper.

Obviously I'm going to test at stock first, and I'll be checking out the overclocking options later on, as it stands though as Neil mentioned he had some issues getting the 10.8's to work with the machine, so I'll be checking them out first to see if they play nicer with a fresh windows install, and I'll revert to the clevo release if there are any issues.

In terms of temperatures, I promise I will be checking that out if it becomes an issue, I have a tube of MX3 and various paste removal sets lying around for just this sort of situation

Akimox - I will test with the dpc latency tracker after stability checking is done, although I can guarantee I have never seen DPC issues on any of my X8100s.

 

Thanks for the link.

I installed the 1.00.11OC bios and now I can run the OCCT GPU stress without graphics issues but in the games concerned, the problem is still present.

I think it's a software problem, just have to wait the news drivers...

Also I have more options in the bios.

 

So these Clevo's also have a proprietary vbios like Dell it seems. It's a shame that nobody has been able to develop a mutligpu solution that can use reference drivers. I think the blame ultimately lies with AMD for having such an awful mobile driver program. That's why I like nVidia so much better, their mobile driver program is light years ahead of ATi in terms of support for reference hardware. Too bad nVidia doesn't have an x86 license.

 

I watched a 3dmark run whilst I was with Kobalt. There's a slight 'left/right' flicker now and then (as though the center of the screen alters slightly for a split second), but the actual rendered graphics were perfect, obviously this is 3dmark so it's running with the video scaled rather than running in native which may fix this. (I've seen this now on Clevo's drivers and the 10.8s) Given the guys saw some anomalies like the screen rolling on the old bios, but not the slightly later revision oc one, I'd like to think a bios revision or so and drivers will fix any issues. It's likely crossfire support is still being implemented and tweaked by Clevo given you often also notice screen and graphics fixes coming up in update notes for clevo machines with NVIDIA cards.
Nearly every manufacturer tags the vbios for their machines so I'm not sure how that comes into it yet.

Given the difference between the system bios variants it could be the fix lies there and not with the GPUs/VBIOS/drivers per se.

 

Funny I didn't pick that up at the time but I was doing those benchmarks in a bit of a hurry and not at the same time as the testing. I was running superpi, 3dmark06, primes on all cores and only one core to record threaded turbo boost on and off CPU oc for varying amounts of time. I expect the max temps recorded in realtemp are from prime runs prior to the 3mark runs and because I was running prime on only 1-2 threads it would have pushed that core up.

 

I can guarantee you that what you're seeing is not a driver issue since I don't have that problem with my M17x-R2 using either Dell's driver or AMD's 10.8 (we can use the reference 10.8 on our system with the latest vbios). Every benchmark runs flawlessly and there's no screen jittering. Definitely seems like Clevo still has kinks to iron out with their 5870m Crossfire setup. Fortunately, Dell's managed to achieve that finally, hope it doesn't take too long for you guys to catch up.

 

By the way I haven't had opportunity to run the benchmarks yet but 10.8 reference drivers installed with no hassle on a clean Windows install, so hopefully it really just is a case of waiting for the next bios revision.
Will see what happens when I run through some benches this evening.

 

Well where to start.
'That voice':Good news everybody! Thanks to a discussion with Pman, I can confirm the overclock BIOS DOES work with 940s, but not with a 920. Overclock settings will hold the core at a higher multiplier than stock with the 940, but not a 920. Please note I'm running my machine stock though, as I've no need for the overclock, nor do I want to open a can of worms, damage my machine or upset Kobalt! But there you go, there is now an overclocking BIOS we can use, don't ask me whether its stable or not overclocked, I'm not going to be doing the testing, nor do I plan to overclock my machine.
<Looking at my luck with laptops you can probably understand why I don't want to add any possible cause of bad luck slapping me round the back of the head again!>

Right onto the more meat and veg of the post.
Basically whilst the machine is not as bad as it would have been by all accounts with the stock .11 BIOS, Clevo still have work to do to get the machine up to an acceptable level with Crossfire. (Which means it works properly aside from driver related issues)
That said I entirely believe the issue is BIOS not Driver related, as every single issue is more reminiscent of an issue with the way the 5870s/mobo interacts with the H06 screen, than an actual driver/gpu fault (no distorted textures, polygons, flickering textures, artifacts etc actual graphical output appears perfect), further reinforced by the drop in issues between the original stock .11 and the later revision .11oc BIOS.

Under both Clevo stock and 10.8, I noticed the aforementioned intermittant graphical waver in virtually every title, if you've ever played electronic pinball its as though you've pressed nudge, and the screen shakes left and right very quickly. Not unliveable but a little frustrating, and it shouldn't been there.

Additionally during feature test 1 of Vantage, I noticed some lines across the screen for a split second, rare intermittant rolling lines during the RE5/DMC4 benchmarks and larger rolling bars intermittantly during the SF4 benchmark. Notice this rolling wasn't as serious as the earlier video in the thread of 3dmark going round and round like a monitor and sync, but, you know, it shouldn't happen at all, so thats not much of an excuse.

With some revision of the BIOS, I feel fairly confident Clevo should be able to resolve this, and providing this is within a reasonable timeframe I will be happy to wait.

Performance wise though, this machine is **** hot, and certainly lives up to the powerhouse laptop status. Everything I have benched is VERY playable at 1080p with maxxed AA, with the exception of Heaven Benchmark in DX11 (but that takes down multigrand desktops, so what can we expect!)

Benchmarks (All run with the CPU at stock!):
Resident Evil 5 Variable Benchmark, 1080p, DX10, 8xAA, all settings on max



Resident Evil 5 Fixed Benchmark, 1080p, DX10, 8xAA, all settings on max



Devil May Cry 4 Benchmark 1080p, DX10, 8xAA, all settings on max



The Last Remnant Benchmark 1080p



Street Fighter 4 Benchmark, 1080p, 8xAA, 16xAF, all settings on max



Street Fighter 4 Benchmark, 1080p, all settings default



Unigene Heaven Benchmark settings and result




3DMark06 1.2



3DMark Vantage (latest version, can't remember the revision no)



Cinebench R10

 

So... who's have trouble with their GameMate Media Player HDMI interface. I seem to have only intermittent luck being able to play X-Box on it (30" 2560x1600 Gateway in shop). I can vary from 0-6 FPS to a full 59 FPS and there is no rhyme or reason as to when it does it. It happened on both sets of NVIDIA drivers from Sager. Happens with SLI on or off, PhysX on or off, have tweaked all setting in Nvidia control panel twice, NO GO. Sometimes it works, others it doesn't. Have updated to software that is on the Sager site to 1.20.2......Looking for clues!

 

@Alex,
Could you also look into the latency while testing?
Congrats on the beast, BTW!
Looks very powerful, on par with M17X-R2 and second to no notebook yet (x7200 ).

 

Yep, if Clevo resolve these issues I'll be very happy. I mean DX10, 8xAA and average over 100fps. A year ago that would have been unthinkable lol

 

Also, could you shrink/attach bench images? Making the page formatting wonky

 

making me work I dont like you anymore lol

Ok gimme 2 mins...oh and Akimox, just for you:



As you can see not a problem whatsoever, thats after scrolling, multiple tabs, a few programmes opened and multiple internet explorer instances

And I've shrunk the previous pictures so as not to kill the forum too

 

It looks like the default TDP is set lower than the M17x-R2. The vantage score is about 1500 points below what I get with my 940xm. That can be easily remedied by employing ThrottleStop and raising your default multipliers + TDP. I agree that the screen jitter/rolling lines issue you are seeing is related to the system bios. Although we never had such an issue with the M17x-R2, we did have the rolling lines problem with the default vbios/driver but ultimately it proved to be a driver problem (pre-Catalyst 10.7). With the latest vbios, we were able to move to the latest Catalyst release and resolve all outstanding issues the system had.

Since you are able to use Cat 10.8, I'm guessing Clevo has an up to date vbios from AMD and now just needs to sort out it's own system bios problems.

 

Aye, obviously disabling crossfire isn't anything more than a temporary reprieve (Clevo have more experience with single cards so it'd be... dissapointing if that didn't work yet either!).
At the moment any score differences between the two machines I'm happy to let sit until the bigger issues are resolved, that overclocking is back in the game means obviously the potential is there for both these machines to be pretty neck and neck.
As you say the issues are seen with both 10.8 and the Clevo stock 8.750 driver, which means our source of issues are different to yours, likely the system BIOS, obviously its possible it could be something else, but that seems the most likely cause, especially given the differences in behaviour shown in .11 and .11oc.

On the plus side no DPC latency issues whatsoever. Loaded SF4 bench earlier with the checker in the background and it didnt even hit 2k. Makes a change from my memories of the M980 haha
Seems the downclocking works correctly with the crossfire 5870's vbios as well, sat there with GPU-Z opened and noticed the clock speeds fluctuatiing dependant on work load and 2d/3d/accelleration state, and temperatures in the sub 50 degree region idle

 

It's a shame these issues weren't sorted before the Crossfire solution was released on the X8100. I hope they do get it sorted for you and don't end up abandoning it in favor of SLi. Clevo has traditionally been an nVidia supporting ODM so that could prove a hurdle in finding a timely resolution. Guess we'll know in a few months but personally I'd return the system and just wait it out instead of being stuck with something that doesn't work out of the box. The X7200 is available now so you could grab a 480M single system for now and grab the 2nd later or just wait on the 470M SLi option which is what I would opt for.

 

I don't disagree that it'd have been better if Clevo had tested it slightly more thoroughly to work out any/all kinks before release, my own patience on the matter being resolved is finite, but after everything thats happened I do really need a machine now (the whole situation and general lack of machine/access beyond my phone I've had since the AW is a large part of the reason I'm no longer on the X32I team for example).

I LIKE the x8100, and Kobalt have done all they can within thier own constraints (seriously cheers guys, you may not always realise it but I do appreciate having dealt with you not Dell!), and I'm willing to give it a bit of time to see if the issue is given the attention it requires.

Once these issues vanish, the machine is quite frankly the perfect (well ok throw in a normal keyboard) laptop for me, it's a decent DTR, without being quite as chunky as the X7200 or the likes of the HDX Dragon, slightly more understated than the M17xR2, but massively more powerful than the average laptop.

It is so very close to being there. (hell even the 5870m downclocking actually works which it doesnt seem to in most forms!)

 

Kobalt should ask Clevo when the next expected bios is. It would be a shame to be stuck with a machine that never gets a proper bios to resolve your problems. How long is your return period? And yeah that offset keyboard would be a little hard to deal with. Why not grab an X7200 though? It's a bit thicker but it's a beast.

 

Yeah, if I dont hear anything I'll be speaking to them in the near future. Return period is debateable, but given my history I don't think thats an issue, I don't want it to reach that point though. The offset keyboard doesnt actually bother me, I just don't like mac style keys, I prefer normal haha.

X7200 - cost/weight etc. For what I want the X8100 is almost perfect, I just need to get these niggles resolved, as if I have to disable crossfire I may as well have gone for a single GPU machine haha

 

If you look at the actual issue it isn't prevalent in all games and benchmarks, see this post earlier:

The BF2 BC testing we carried out (which was one of the games that was completely unplayable before we shipped any of the X-Fire systems out) ran absolutely perfectly by the time we had updated the BIOS and played with drivers.

It's not going to be stuck without a full fix - progress has already been made.

 

Aye, the 'pinball nudge' effect I see in everything is not game-ending glitch although slightly annoying, and I'm sure will go with time, I'm sure Clevo are working on the larger issues as we speak (well not perhaps as we speak because they're probably sleeping/off work haha).

I just really hope they make releaseable progress in the form of a .12/13 soon

Unigene Heaven is another program that runs flawlessly (although admittedly thats a tech demo).

 

here, run this one...it has already set presets.
http://91.121.148.119/downloads/hwb...en-Benchmark-Application-1.0.2-Setup-full.exe

 

Gimme 20 mins to download it and I will. Will uninstall normal heaven as well just incase it causes any issues.

 

its settings are preset.
you can run dx9 and dx11

and you cant change anything. just click dx9 or dx11. then check your score at the end. get a few more people in and you have a data base going...

 

Should only be about another 5 mins til its done

 

Ok, submitted under same username as my NBR name. (100% stock clocks etc across the board)
Basic was 1110, and Xtreme was 668

 

You call "rare flickering" and Alex's "everything has screen jitters" absolutely perfect? It's not running absolutely perfect if the image is not properly presented on the screen. You should give your customers a firm ETA on when they should expect a fixed bios or at least a statement from Clevo on their intent to fix it. Hell, we get more out of Dell reps by pushing them around on NBR. Fact is you and other vendors are shipping broken systems to customers, end of story. I'm sick of companies thinking it's fine to ship expensive products to customers that don't work properly out of the box (this applies to Asus/Dell/Clevo).

 

nice.

and how did the screens look?

 

Even though I usually start pointless and stupid arguments with you,...you're mostly right.
And since you managed to make your system work properly (after much pushing and bashing), there are no excuses (like "SLI/CF isn't perfect") for any company.
The systems should work flawlessly out of the box, and not 1 year later.
That's why I'm so reluctant to approach small companies and resellers. Dell and HP at least will try harder to solve the issues and keep you satisfied (though their progress is sloooooooow).

@Alex,
Could you make a pic/video of the "jittering" that you see?

 

No, I said nothing of the sort, you have completely misquoted me and are putting words in my mouth.

Please read my post, I was specifically refering to BF2 BC and stating that that was one of the games that other customers (of other companies, not Kobalt) had had serious issues with and that we had seen a 100% improvement i.e. after updates it runs flawlessly.

We actually held off shipping them until after the X-Fire setup was available elsewhere because of some issues we picked up on under our testing. At the same time we can't test the setup with every game available and we haven't had reported problems from other customers, though to be honest we have only shipped a very small number of X-Fire X8100s. Now that we know that there is still an issue we can deal with it by working with the relavent parties.

Alex came round to pick up his laptop from our office last week, he is a customer we know very well. He was in our office for the best part of two hours and I was chatting with him for a good 45 minutes. As with all our customers we will bend over backwards to make sure all reported issues are dealt with ASAP.....just ask anyone.

I explained above that we delayed shipping the X-Fire X8100 because of an issue we found under testing. We delayed because of our findings and specifically because we DON'T think it's fine to ship products that aren't working properly out of the box. Alex knows this because he was kept fully informed during the build.

Because you seem to have absolutely no experience of our company and are drawing wild conclusions I will give you an example. I spent a huge amount of time working on the 980NU audio issue and through the amount of information we passed to Clevo and Nvidia we implimented a fix that solved almost all our customers reports. It wasn't reported to any company until about 6 months after the first unit shipped, after it was reported to us we knew what to look for and added this to the system testing procedure. In fact only about 5% of our customers experienced any audio issues, though that's 5% too many. The customers that didn't find the fix workable we accomodated at our financial expense.

Aikimox - we might not be the size of Dell but if you're seriously suggesting that I don't "try hard" to solve issues then you clearly don't know me at all. The above example is just one of many. I don't know where the 1 year fix came from but progress has already been made and I'm sure that it will be sorted quickly. It makes no sense to complain about me when I'm getting the issue sorted out, especially when you don't even own the laptop in question but Alienware machines. Something else I think you need to appreciate is that we are not a drop ship reseller, we are a system integrator which means that we build, test and directly support all of our products. This is wether they are laptops, £10k desktops, workstations, render farms running in 4D studios and so on. I work with contacts in major brand component companies in both the mobile and desktop sectors and quite often, because of the nature of the systems we build, we come across issues that more mainstream SIs don't.


If other guys have had issues with their X-Fire X8100s then they need to tell their SI. If the issue is being reported to the right parties then it will be fixed.

 

Neil,
I wasn't talking about you specifically.
But as it happens, I was among the unlucky M17X-R1 owners exactly 1 year ago and shared the same fate with Alex and all the rest who had Clevo M98nu's.
Guess what, many months after our R1 issues were resolved (by switching to 4870's, replacements, refunds, etc), Alex and many others were still bashing the walls.
It's very simple. Companies like Dell have more options to deal with the problems directly. They just have more resources. Same as they can offer a NBD warranty, while you can't. And I understand your feelings very well, since I'm in a similar boat ***cough*** NDA...***cough***
You do your best, of that I have very little doubt, but there's a limit of options which is a logical thing.
And "try harder" is just a figurative expression, the meaning is resources (don't exclude the fact, however, that not every reseller takes this to the same level as you do).
We can do a great job, but if the key players (Nvidia, AMD, Clevo) screw up like back then, - we just do what we can, wait and desperately bash the walls...
That's all, no offense.

 

First, my apologies if I misquoted you. Second, it's great that you go to such lengths for your customers and that's what every vendor should strive to do. However, I don't see how you could have missed the screen jitters that Alex describes if you put the machine through heavy testing as claimed? I know if I ran a company that sold boutique systems, I'd have at least the top 5 popular PC games on hand to test as well as Vantage and a few other synthetic benchmarks. I'm sure with all the orders you get, time constraints don't allow for hours and hours of testing, but a streamlined 2 hour testing session for each custom built isn't out of the question--I think even Dell manages more than that on each AW they sell.

But my point stands, now that you are aware of the issue, you should at the very least advise customers on when they can expect a fix and whether Clevo has commented on it or not. Because as of right now Crossfire does not work properly for the X8100. It would be a shame for customers to buy this system in good faith and then 6 months later be stuck with the same problem while Clevo focuses all it's attention on it's newest darling the X7200.