Clevo Overclocker's Lounge

http://www.hardwarecanucks.com/foru...vidia-titan-x-12gb-performance-review-18.html

37% in DX11, 35% in DX12.

And yeah, I can get comparison benchmarks between titan x and 1080 ftw in my ultrawide res.

I almost wish I went with an OLED 4k screen/tv.

 

From the reviews I've seen, even an overclocked Titan X (P) won't be able to do 4K @ 60fps on ultra on every game. However, it comes close and plays most games fine. 1080 SLI is definitely the minimum for 4K 60fps assuming the game supports SLI otherwise a single Titan X (P) will do the job fine.

 

I like the ultra-wide screen too myself, either 1440p/2160p, but maybe a bit bigger than the current largest, 35"? I'd like a little more height.

When I put up projection screens, I like to make them extra long to get a "big" 4:3 and roll them up a tad to make the height match the length for 16:10/16:9 and 21:9, all in one. Multiple settings in the controller select the length and set the drop brake at the right stops. I try to use the entire height of the ceiling to floor, and of course side to side for screen material.

I like that HWC did 1440p and 4k both for comparisons, here are their DX11 / DX12 result tables:

Yup, you can cover all the bases by using a Titan X(P) SLI

 

I ran my Titan-x at clocks to make the most out of it (1485/8000) under water and properly gain over the 980 and 980Ti but the card failed after a year, it was not designed to run at that (the card).

So yes the new one can go to above 2ghz but for how long?

 

So far this is all on stock vbios, no need to worry.

 

Replaced over five years old thermal paste on i7-920Xm - My antique Clevo W860CU (i7-920Xm + Gtx460M) to the better Grizzly Conductonaut (I'm empty for CLU). A Nice reduction in temperature(Up to 13C degrees reduction in max temp ). I tested Cinebench R11.5 and Wprime 1024M stress test before and after Liquid metal and with a slight overclock @2666MHz on processor.
Cinebench R11.5 old paste 87C and with Liquid Metal 74C = 13C degrees lower temp.
Wprime 1024M old paste + throttling 95C and with Liquid metal zero throttling and 84C = 11C degrees lower temp.
After new paste there is only 1 C degrees difference between the cores in bench Old paste have up to 11C degrees difference between coldest vs hotest core!!

Shall post a few bench with 3.2GHz later.

Grizzly Conductonaut(sorry for the bad image )


Old paste. I don't remember the brand


Liquid metal


[email protected] (Clevo W860CU)
Cinebench R15 - 497cb http://hwbot.org/submission/3281877_papusan_cinebench___r15_core_i7_920xm_497_cb


Cinebench R11.5 - 5.49 pts http://hwbot.org/submission/3281823_papusan_cinebench___r11.5_core_i7_920xm_5.49_points

Stock i7-920Xm Notebokcheck


Edit. I tested 3Dm11 benchmark now with @3.2GHz. Crash!! The too weak 120W psu did not manage the task. It failed!! I can fry eggs on top of this psu now, HaHa. Why deliver Clevo W860CU with a all too weak psu? Have Mr. Azor worked with Clevo around 2009/11? This is actually not the first revision of AW "Echo" with Hybrid bios, LOL
Overclocked i7-920Xm and stock 460M. I can try with ice on top of the psu, when I get time

 

That does not sound like a stock PSU for a unit that shipped with an extreme edition.

 

Oh yes. Multicom Norway shipped this psu with the setup ^^^. I have only put in the 460M. See image of psu. 20V@6A


What psu can I use instead? Have you part number for a replacement?

 

If it's a barrel tip use a delta 180w with a 5.5x2.5mm tip and 19 or 19.5v output.

 

19.5 vs 20v is ok? I know the opposite isn't ok.
I have a spare Delta 19.5V@240w but with the thicker barrel tip intended for AW. Can I just use the cable from the 120w I have and use it on the 19.5V@240w? What about the id chips Dell use?

 

Be careful when mixing brands / cabling of power supplies at that current range. Be ready to pull wires, and keep a fire extinguisher handy.

"Don't cross the streams!!" - Egon

 

1v either way is not an issue. Replacing tips/cables is more involved and 60w extra should be enough for now.

 

I push 175w in 3DM11 physics test. Combined test push only 160w and the other graphis tests right below 150w. Never seen physics use more power from psu than the other 3DM11 tests, LOL

With ice bag on the psu, I fixed 2290P in 3DM11 without psu crashed http://www.3dmark.com/3dm11/11475223
I need only 73P for top 1 valid score in 3DM11 with 920xm/460M

 

That from the wall would leave plenty of room on a 180w brick.

 

@joluke - this is for you, bro. Here's a new 880M SLI world record. Once 3DMark adds the hardware ID (requested) the score will be validated with the GPUs recognized.

http://www.3dmark.com/fs/9676618

 

nice to see the 880Ms working PnP in the phoenix have u also tried 780M or even older yet? u could potentially break a looot of world records that way!

Sent from my Huawei Mate 8 NXT-AL10

 

I don't own 780M cards any more.

 

Nice. What about Max gpu temp? 880m is known as a hot card.

 

Mine in current weather (plain summer with 40 degrees Celsius) is running at 75 degrees (Celsius). With Fn + 1 fans. If I dont push my fans to maximum I will get temps of 83 degrees which isn't that bad with 2 hours of gaming and thermal paste Arctic mx 4

 

Here's a new 880M SLI world record in Fire Strike. Once 3DMark adds the hardware ID (requested) the score will be validated with the GPUs recognized.

http://www.3dmark.com/fs/9676618



Here is the highest recorded 880M SLI score for 3DMark 11 on a laptop with a quad core... Not far behind the highest score recorded by a P570WM with 4930K CPU.

http://www.3dmark.com/3dm11/11477144

 

So drivers were a big problem with overclocking. As suspected.

 

Definitely. I think we have both known that all along and could never prove it beyond the shadow of a doubt. If I had a better vBIOS for these 880M cards I might come very close to 980M SLI. (Further confirmation of NVIDIA manipulation shenanigans.)

Here's a new world record for Vantage... even beat P570WM with 4930K and 880M SLI, LOL.

http://www.3dmark.com/3dmv/5494845



@D2 Ultima - more proof for you, bro. See my post prior to this one as well. 880M SLI finally performing right now that NVIDIA accidentally released a cancer-free driver. I don't have any way to check if the same driver fixed the Alienware 780M SLI throttling issues. These are stock Dell 880M cards out of an Alienware 18 with the basic @svl7 / @johnksss public vBIOS. With more voltage and higher power target slider limits (currently limited to 160, needs to be about 225) the scores would likely be much higher and I'm pretty sure I could go a good ways past 1200 on core with 1.200V. The P870DM-G pulled 520W with this overclock according to my UPS meter.

 

Hummm, that's a new version of AM3

 

I just took first place and set new world records on Fire Strike, 3DMark 11 and Vantage for 880M SLI... These comparisons links are against current first place scores. Once 3DMark adds my hardware ID these will show in search results and displace the guys currently in the #1 spot to the #2 and #3 spots based on what I have uploaded today. To my surprise, the top scores I beat were a P570WM with 4930K and 880M SLI.

http://www.3dmark.com/compare/fs/2402025/fs/9686855

http://www.3dmark.com/compare/3dm11/8510683/3dm11/11477498

http://www.3dmark.com/compare/3dmv/5067708/3dmv/5494902

 

Interesting comparison... 980M SLI (overclocked, but not top scores) and the above 880M SLI 3DMark 11 run, all three in quad core laptops. Not as good as 980M SLI, but getting closer than we might have expected.

http://www.3dmark.com/compare/3dm11/9171430/3dm11/10102817/3dm11/11477498

 

Not surprising as a large part of maxwell is the improved memory technology, something that benchmark ignores.

 

I've got the memory overclocked higher on these 880M cards than most people are capable of overclocking their 980M cards. And, yeah you're right it makes less of a difference in 3DMark 11 scoring. But, an unstable memory overclock will fail in 3DMark 11 whereas it might actually get a pass on Fire Strike, since it seems to be far more forgiving in terms of stability. That's probably a good thing since most noobs seem to favor it.

 

All benchmarks do not test the exact same as specs as each other. That's why they have a variety of them....Just something to keep in mind. Like for example, Timespy. It will drop an unstable memory overclock pretty darn quick. And without you seeing the blink dots on the screen.

Is that why you would skip 3dmark11 when you had your super secret voltmod going on? Back when I told you that would not work on 3dmark11 because it's more of a core bench.

 

Indeed. Time Spy has a very low tolerance--almost none, in fact--for an unstable CPU, GPU core, GPU memory or system memory overclock. It will crash almost immediately if anything is sketchy or borderline.

 

Due to memory compression however the 980m is moving a lot more data for the same speed.

 

My 880M back in the day runs...

 

Yeah I would get up to around 1700mhz on my 780m so certainly higher clocks than my 980M in raw Mhz.

The cramped 980m pcb was more limiting than the memory or IMC I found.

EDIT: 1750 Mhz even

http://www.3dmark.com/fs/1407660

 

Yes, but what about everything else?

 

Be sure to disable the "Systeminfo" feature in AM3 before running the benchmark or the file you save to the desktop to upload at HWBOT.org will be corrupted and unparseable. I did not do at first and I cannot upload my highest score (430509) in that screen shot previously posted, so I can only take credit for my second highest score.

 

I was able to submit without altering anything.

 

I had trouble with the AM3 dump files created on both of my machines and found on HWBOT forum where that was a known issue on some systems. The solution was to turn off the SystemInfo feature and that fixed it for me.

Did you run it under W7 or W10? If W7, maybe it's a W10 glitch.

 

windows 10

 

Hmm... I wonder if maybe something with the DDR4 memory. When I run SystemInfo enabled it shows the memory as "out of spec" because it does not recognize the DDR4.

I will try it again later with the P750ZM and see if it works with the DDR3.

 

Maybe. Ill run it on my desktop since that has ddr4 3600 mhz ram in it and see if it has an issue as well.

 

3DMark added the 880M hardware ID and all of the #1 rankings for 880M SLI are now validated.

http://www.3dmark.com/compare/fs/2402025/fs/9686855

http://www.3dmark.com/compare/3dm11/8510683/3dm11/11477498

http://www.3dmark.com/compare/3dmv/5067708/3dmv/5494902

 

Finally on top for i7-920xm/Gtx460M single graphics in 3DM11. Valid results 2420P with the old lady Clevo W860CU. The ice bag helped to cool the weak 120w psu
http://www.3dmark.com/3dm11/11479823

 

New better 3DM11 record - 2454P from the old lady http://www.3dmark.com/3dm11/11479949



Vs second place Valid results 920xm/Gtx460M


New a slight better 3DM11 - 2458P http://www.3dmark.com/3dm11/11479980
See Cpu/Gpu temp(Liquid Ultra )




Edit. New better 3DM11 - 2459P http://www.3dmark.com/3dm11/11480018

Not a record score this time in Cinebench R11.5 but the 7 year old antique [email protected] GHz love power . Stock voltage and over 86w in bench. Maximum temperature is great when you use Liquid Metal . These new slim modern AW Turdbooks would struggled with load at this level, LOL​

 

Firestrike ultra

http://www.3dmark.com/fs/1434239

4.4Ghz 4920k CPU and the 780m cards were at 1171/1750 ram was 4x4gb at 2133mhz 10-10-10-23 1t.

3dmark11

http://www.3dmark.com/3dm11/7721191

Same as above but gpus at 1151/1750

 

Ah, thanks for that.

 

No problem I'm looking forward to properly joining in again.

 

Not sure if anyone is up on these, but these are totally awesome!
http://www.geek.com/news/lian-li-an...d-computer-cases-that-are-also-desks-1625712/
Never even peeped them out last year, but saw one in person. Very nice and will compliment anyone's decor. They are building the big one for someone at my Microcenter.

 

Sky Diver 36246P http://www.3dmark.com/3dm/14005168
220/500

 

Did that driver make a difference?
Might be lower gpu score, but still king of the hill.
http://www.3dmark.com/sd/4221549

 

Nice score I have not tested out any new drivers for this machine since before my month-long vacation at the Canary Islands. That means before 28 June . I have worked with my wife's old lady above ↑↑↑ The old lady got some shine up(Liquid metal ) + testing afterwards.