Lenovo ThinkPad X1 Extreme benchmarks + temps

Those of you who check the Dell XPS forums know that I have had a couple of XPS systems (multiple from every model) and have done some benchmark/temp mini-reviews on them. I received the X1 Extreme a few days ago and decided to do the same for this. This is a "live" review as I will be updating it over the next couple of days. I welcome any benchmark requests (as long as I have it or it is free that is, lol).

Specs of my X1 Extreme:
CPU: i7-8750H
GPU: 1050 Ti Max-Q
Ram: 16GB DDR4 2666MHz (single stick)
SSD: 512GB Intel 7600P (SSDPEKKF512G8L)
Screen: FHD

Stock Benchmarks with windows power slider at "better peformance" and Lenovo Vantage left on "enable intelligent cooling"

Time Spy - 2402
Time Spy Extreme - 1106
Fire Strike - 6549
Fire Strike Extreme - 3324
Fire Strike Ultra - 1694
Sky Diver - 19280
Cloud Gate - 24046
Ice Storm - 79560
Ice Storm Unlimited - 160401
3DMark 11 Performance - 8661
3DMark 11 Extreme - 3262
Unigine Heaven Extreme -1113
RealBench System Score - 91,862
CPU-Z Multi Score - 2611.3
Cinebench R15 - Open GL 96.16FPS, CPU 915 cb, Single 174 cb
Corona - 4:08
7-Zip Total Rating - 32880 MIPS
PCMark 10 - 4892
PCMark 8 Home Accelerated - 3672
PCMark 8 Creative Accelerated - 3516
PCMark 8 Work Accelerated - 5294
Y-Cruncher 25M (sec) - 27.068
Y-Cruncher 1B (sec) - 133.006
Super Pi 1M (sec) - 9.248s
Super Pi 32M (min:sec) - 8m 48.422s
Prime95 64k Bench (iter/sec) - 20068.76
Prime95 2048k Bench (iter/sec) - 210.53
AIDA64 Cache -

Thermals
Testing for the thermal section is done with a variety of tools. I am using Prime95 29.4B7 to test the CPU, RTHDRIBL as a light GPU load, and RealBench 2.56 as a heavy workload on both the CPU and GPU.




Undervolting
The CPU in my little X1 seems to be pretty decent quality wise. I was able to undervolt it by 165mv and it is 4 hour RealBench, 10 hour Prime95 Small FFT, and 8 hour AIDA64 system stress test stable. I am not concentrating on a how to for this but can add more detail if anyone requests it.

Repasting
Some how Lenovo managed to apply more paste to the X1 Extreme than Dell did their XPS 9570, no small feat mind you. Removing the stock paste and applying some Kryonaut yielded the results above. Not a huge decrease, but worth it in my opinion. I may try something different like ICDiamond to see if it helps because the dual tripod design of the heatsink doesn't apply much pressure.

 

Modifications
So I think I've arrived at a point where I am fairly happy with my results finally. Here is a list of everything that was done before I re-ran all of the benchmarks.
1. Undervolted the CPU and CPU cache by 165mv - by far the best "bang for the buck" so to speak.
2. Repasted the CPU and GPU with Kryonaut - worth it if you're comfortable with repasting and already have the supplies
3. Set the Thinkpad Vantage Program to optimize for performance - seems to increase the long and short turbo boost to a higher wattage. Causes the system to run hotter but not out of hand.
4. Removed the fan grill "sticker" from the bottom of the case - very much worth it to me as the noise these fans put out were beyond annoying prior to doing this.
5. Overclocked the GPU by 200MHz and the vram by 350MHz - not only do I get a big performance boost, it actually runs at a much lower VID.

Benchmarks with the above mods:
Time Spy - 2549
Time Spy Extreme - 1184
Fire Strike - 6942
Fire Strike Extreme - 3522
Fire Strike Ultra - 1795
Sky Diver - 20829
Cloud Gate - 27566
Ice Storm - 76172
Ice Storm Unlimited - 161975
3DMark 11 Performance - 9242
https://www.3dmark.com/3dm11/130691973DMark 11 Extreme - 3455
Unigine Heaven Extreme - 1191
RealBench System Score - 114,708
CPU-Z Multi Score - 3428.2
Cinebench R15 - OpenGL 101.61FPS, CPU 1206 cb, Single 171 cb
Corona - 3:07
7-Zip Total Rating - 38846
PCMark 10 - 5224
PCMark 8 Home Accelerated - 3524
PCMark 8 Creative Accelerated - 3492
PCMark 8 Work Accelerated - 5191
Y-Cruncher 25M (sec) - 25.041
Y-Cruncher 1B (sec) - 121.208
Super Pi 1M (sec) - 9.264
Super Pi 32M (min:sec) - 8:45.5
Prime95 64k Bench (iter/sec) - 27908.59
Prime95 2048k Bench (iter/sec) - 241.88
AIDA64 Cache -

To make things easier to read here is a before and after:

 

Here is a quick and dirty against my Dell XPS 9570 benchmarks and temps both stock and modified:



Obviously there are pros/cons to each laptop. I would say the X1 feels much lighter and easier to carry around, has a better keyboard, has a spare NVMe slot, and doesn't have the GPU throttling that the Dell does. The XPS feels better built to me, the screen is much better, much better battery life, and is much faster at CPU related tasks than the X1.

So after reading about @B0B's experiences with the Overpowered laptop and having single vs dual channel memory, I decided to purchase 32GB of matched memory to re-run some of the benchmarks. There are some decent gains to be had even from a synthetic benchmark point of view and it helped to even out with the XPS by a bit:

 

There is lot of talk in reddit about somekind of "mesh" in the back panel that interferes with the fans. Many have said that removing the mesh will help a lot with the fan noise. Not sure what mesh they are talking about as i haven't seen it in any pictures.

 

I was actually looking for a replacement back panel so I could cut out the mesh. I will try to take a picture of it, basically it's just the fan grill that covers the fan, the design is silly for sure. I sure hope it helps with the fans because that (and the screen) are really the big things holding me back from keeping it...

*on edit* after looking again, it's just a stuck on mesh so I went ahead and removed it. The fan pitch is much much better now. It's still louder than my XPS but not near as annoying anymore!

 

Alright I have the benchmarks with the "Mods" in place. I will finish up my excel calculations later today and post everything up, including the comparison to my Dell XPS 9570...

 

Alright that about does it. Let me know if there are any benches or anything someone would like to see. I am still undecided which laptop I will keep as they both have their ups and downs. I will play with the lenovo some more though.

 

Good job, Man! That's a lot of hard work leading to amazing results. I was looking at this model but eventually gave up on it. Too many compromises in the end. Entry level GPU. Limited storage for a 15" machine. You absolutely have to undervolt if you need real performance. And those temps are still not quite convincing. If your room temp is 25-30C, it will easily break 90C even with repaste + undervolt + raised back. No black magic here. But it's a very nice machine for sure. Very stylish and fits the niche quite nicely.

Personally, I ended up getting a E485 with Ryzen 2700u + Vega 10. Only cost me 500$, but those 10W can provide half of the CPU+GPU performance of a X1 Extreme for less than 1/4 of the cost. In addition, I have a crazy storage capacity in this slim 14" package - Full-size NVMe + 2.5" bay = 6TB SSD storage that can be achieved right now. Finally can happily ditch W10 and move to the dark side, as linux seems to work very smooth with Ryzens.

 

Yeah I am fortunate to have an office in the basement so my temps are pretty much a constant 22C. I'm actually pretty surprised by the temps of the laptop. Lenovo does limit CPU power pretty quickly so that helps with the temps. The 1050ti is good enough for the games that I play when I'm not at home on my 2080ti and 34" monitor. Sadly there aren't any great thin gaming laptops aside from the GS65 and Blade 15, but those laptops are $$$. I paid $1200 for the X1 and $1000 for the 9570, but the 9570 was refurbished.

 

Yeah, I agree, not much choice when it comes to thin gaming laptops. Nothing even remotely close to my 2 year old OLED AW 13 + Amplifier. The new AW M15 has terrible cooling system. Hope they reintroduce OLEDs next year along with better cooled laptops.

 

Thanks for sharing this !

 

You're welcome, it was fun. I hope it helps someone who is thinking about buying an X1 or XPS.

 

Are you going to try and overclock the display? If you try, can you post your results?

 

I haven't planned on it, but can give it a try if you'd like. Are you hoping to hit any particular Hz?

 

Just curious as to how far you can reasonably push it. Any Hz increase is nice, but getting 75+Hz would be pretty cool. If it could reach 90+Hz, that would be amazing

 

Sadly I don't think it's going to overclock at all. Trying to increase the refresh rate by any amount results in a pixel clock that exceeds the bandwidth.

On a slightly more happy note, adding dual channel memory helped out with some of the benchmarks and helped it catch up to the 9570.

 

That sucks. When Lenovo fixes the dGPU only option, I think it would be worth trying again

 

I believe bandwidth is limited by the connection type is it not?

 

Apparently, sometimes it can be because of some strange driver issues. Forcing dGPU to output to the display skips and iGPU and thus, skips the intel display drivers. At least this is the case IIRC

 

You're absolutely right, I enabled just the dGPU and ran it and was able to get up to 67Hz before I started getting frame skipping/garbled screens. Not great but better than nothing I suppose.

 

It seems like you got lucky, most NV156FHM-N61 panels only go up to 65Hz or so. It might be interesting to see if a panel swap can make a 100Hz or so panel work...
Thanks for your results

EDIT: Did they fix the bricking issue already?

 

Sadly I have no idea about the bricking issue, lol.

 

https://www.reddit.com/r/thinkpad/comments/a2g0k4/warning_do_not_change_from_hybrid_graphics_to/

 

Well I had no problems switching to and from discrete graphics.

 

You dodged a bullet! I was surprised to see that you flipped it on and it didn't brink your system. I thought you would wait for a newer bios version before testing

 

Hah I guess I just lucked out (had no idea it was an issue). I updated the bios on the 24th and here is what it says:
- Fixed an issue where the system might not boot with Discrete Graphics.

 

Holy, talk about good timing

 

Why did u not undervokt the iGPU?
Any advice on how to test overclocking speeds of the iGPU to get a stable value?

I moved from a XPS to the X1E any do t regret it. Especially the keyboard on the X1E is a big plus.

 

I undervolt the iGPU but not for these benches. To test the undervolt of the igpu I use the nvidia control panel to set a desired application to use the dedicated gpu. I prefer to run loops of 3dmark, but you pick what works best for you. So far I like the X1 but wish the screen was better as well as the battery life. Those two things are major negatives coming from the XPS...

 

How far did you tune it? I helped a friend tune his X1E and could push it to 7.5+ hours or so of web browsing.

 

What did you do to oc the panel? I'm not familiar with this..

 

Hey,
I tried to follow this guide on overclocking an LCD with intel graphics.
Sadly it didn't work out because of "bandwidth" limitations. I then turned on the dGPU only mode in bios and was able to set a custom resolution with the NVIDIA control panel. Sadly not worth it for the extra 7Hz.

 

I wonder if you can drop in a touchscreen into the FHD model. Does anybody know if you can do this?

 

You need the cable with touchscreen data which doesn't exist for the FHD panel.

Also the thickness might be a problem for the slimmer display assembly.

 

Ah, that sucks. 4k touch is overkill IMO. I hope there will be an X1E yoga that sports a 1080p touchscreen with pen support...

 

First thing is to congratulate you for sharing such a laborious and systematic work. I have a X1E manufactured last December and I'm trying to squeeze it to the max. Although stress benchmarking using for hours Prime95 or Aida64 showed that a CPU undervolting of -165 mV was"stable" in my machine capable of significantly boost CPU performance, as shown by CINEBENCH, I'm getting eventual OS freezes while operating in no-so much-stressing conditions like using Microsoft Word on Win 10 Pro with Chrome running in the background. Also, with that level of undervolting, system eventually failed entering in "sleep mode". I've also tried this level of undervolting after repasting with Thermal Grizzly Kryonaut but the problem persists. Finally, I found that my stable "CPU undervolting" was -138,7 mV in any operating condition.
Same deal with GPU 1050 GTX MQ. Synthetic tests showed that +200 MHz GPU/350 MHz vRAM OC was "stable". However, real life test showed that GPU can't be forced beyond 172 MHz while playing games such as "Metro Exodus" or "Far Cry 5".

My point is....did you get "stable" real life operation with the proposed CPU/GPU undervolting/OC values, leaving aside the synthetic tests?

 

Yes my system was totally stable in everyday use and gaming at those settings. Bare in mind that I had an excellent CPU and GPU when looking at my past 5 Dell XPS 9570s. The silicon lotto is real...

 

Yes, I guess it will be a lottery thing. The funny thing is that with my tuned settings I'm getting significantly more FPS in demanding games that those reported elsewhere with laptops sharing same CPU/GPU combo. Do you have (at hand) any FPS measurements in any current game to serve as a reference for my system numbers?
Thanks!

 

Sadly I never logged any frames per second, only temps and frequencies.

 

Ok, I see. Thanks anyway. Overall, I'm pretty happy with the X1E with the exception that I had my MoBo replaced due to a DOA SO-DIMM slot. My only current complaint is battery life. Not sure at all what is it exactly. I'm running 90% of time on A/C with the 80% charging cap. I realized that just after unplugging, that 80% capacity of battery gives me between 16 and 3,5 h (second lowest level of performance in Win 10). I guess monitoring software is not so much polished or maybe battery life is bad. Should I run a battery calibration?

 

Do you mean 1.6 and 3.5 hours? I typically got between 6 to maybe 8 hours on mine, but I didn't limit the charge capacity on the battery. Is there something that is running that is eating battery? Keep in mind when I was on battery it was just using the internet and office applications.

 

I really mean 16 h! Clearly the battery power estimate is working atypically on my system. I'm getting those readings with office and light internet use, brightness at 75%. When I have the time I'll run a full test of battery life from 100%-0% to check it out what is the true behind this issue...

 

I guess I'm confused. You're upset about battery life and you can get 16 hours out of it?

 

No, I think that 16 hr reading is totally unreal. My complaint is that in my system, battery life expectancy reported by windows 10/lenovo tool is never accurate, even after several months of use...

 

Oh I see, I've never listened to the guestimate that windows provides. My 6-8 hours is real world usage with some remaining battery life. I have a few PC Mark 8 battery benches as well. Not as good as my 9560/9570s but good enough for what I did.

 

I am trying to decide between XPS15 and X1E for video editing and animation workload. Also I am going to be in warm room temp.
So I wonder which machine would you recommend? I would go with 4K screen option so my most concerns are
1- performance
2- surface temperature
3- fan noise (I know x1 has a louder fan, but I wonder how bad it is comparing to xps)

 

The X1 isn't louder than the XPS if you remove the fan grill cover, otherwise it is super annoying and loud. Performance wise they are about equal except the X1E favors the GPU under heavy loads while the XPS favors the CPU. I had no issues with surface temps, but I wasn't stressing the CPU/GPU on my lap or anything. Really it's a toss up between the two and depends on what you're looking for.

Here is a quote from my 3rd post when looking at the two:
"Obviously there are pros/cons to each laptop. I would say the X1 feels much lighter and easier to carry around, has a better keyboard, has a spare NVMe slot, and doesn't have the GPU throttling that the Dell does. The XPS feels better built to me, the screen is much better, much better battery life, and is much faster at CPU related tasks than the X1."

 

Any idea how this would stack up vs the MSI PS63?

 

Sadly not personally but based on the specs, the X1E and XPS will be a good deal faster. That PS63 looks nice though.

 

I was thinking of getting one of these or a P1. The P2000 GPU is basically the same, but drops 1GHz on the memory clocks. How much of a difference would that make, and can I OC the memeory to make up for the difference