SACbomber

Doubting there's any meaningful airflow over the CPU/GPU dies and VRMs/VRAMs. All the thermal putty/pads/paste would cut off the airflow.

For those unfamiliar: https://game.intel.com/us/stories/200s-boost-overclocking-profile/ I have an 285HX and XTU lets me increase "SA fabric ratio" which I think is uncore/NGU. Didn't see a D2D option in XTU so that might be in my BIOS. My RAM is 6400 MHz(MT/s) CSODIMM and I wonder how overclockable they are. Has anyone tried this?

Googled "Zero 18 laptop" and went from there. Maybe, except that third fan isn't blowing towards anything meaningful. At least the third fan in the Tongfang blows across one RAM stick and what looks like an M.2 heat shield/spreader.

Think we're far from the Sager/Clevo theme now but here's this laptop's cooling solution: Looks like three fans and five or six heatpipes. Reminds me of ASUS's solution with three fans. What's the third fan meant to do? Cool the VRAM/VRMs?

Nice chunky chassis and proper separate arrow keys. Need to see the cooling solution. Don't like the reacharound USB ports on the back. RGB in the vents is a strange design choice, especially since they're not even full-width according to the renders. Think it looks really strange although some fun can be had if you could program the lights to change according to fan speed/heat output, like jet engine exhausts.

More notes: +400 MHz GPU core is not stable. Had no issues running video tasks or benchmarks/gaming but OCCT's power test would catch an error before 5 minutes, guaranteed. Dropped core to +350 MHz and saw no more errors. CPU wise, running -60 mV on all cores crashed the system at idle. No stress test could crash the system or catch any errors, yet idling with a few tabs open made the system reboot. Thought that was really funny. Dropped my undervolt to -50 mV on all cores and cache. 3D applications would sometimes not make the GPU draw full power for some reason. Boot system, start CP 2077, see 170+ W power draw on GPU, makes sense; close CP 2077 then start it again, only draws 65 W with frames down to 35. Try another 3D application such as another game right afterwards, same issue. Have to reboot to resolve. Seem to have mitigated this issue by setting affected applications to "prefer maximum performance" in nVidia Control Panel's "Manage 3D settings -> Power management mode". Going to call it a driver issue for now. Blackwell drivers have been troublesome since launch. Did check MSI Center, was on Extreme Performance mode all the same.

Decided to upload BIOS files for this Titan since MSi removes old versions from their support site. Just in case someone (like myself...lol) needs them in the future. Downgrading BIOS is possible after unlocking a setting in the advanced BIOS. E1824IMS.112 E1824IMS.113 I'll see about doing that somehow. Might start playing with the voltage curve in Afterburner. More to come.

Would shunt mod my Titan only after the warranty period ends. Not risking anything right now. That core clock is definitely from some sort of mod alright. Disappointed that 3DMark doesn't record the GPU power level. My 5080 temperature never went above 71 degrees and it usually holds around the high 60s while sustaining 170+ W of power draw. There's some headroom here to take advantage of. The Raider looks incredibly gaudy for sure although that image is simulating the keyboard lighting.

+2000 MHz GPU memory, +400 MHz GPU core, Steel Nomad results compared to stock clocks: https://www.3dmark.com/compare/sn/5571886/sn/5571836 Same conditions, Speed Way compared to stock clocks: https://www.3dmark.com/compare/sw/2183503/sw/2183493 Expected more gains honestly. This thing can go faster. Can feel it. Shunt mod is the way.

Rick from the MSi board here, hope you don't mind me intruding... Have you got any benchmarks of your X580 @win32asmguy? CineBench 2024, 3DMark, the usual. Really want to compare with my TItan. Expecting you to have a 10-15% graphics performance advantage over me while maybe slightly behind on CPU/RAM.

+450 MHz GPU core is not stable. CP2077 crashes immediately upon starting benchmark. Trying +400 MHz mem, +400 MHz core. Not done yet. Historically GPU mem has been a lot more overclockable than core. More to come.

+300 MHz GPU memory, +450 MHz GPU core through MSi Afterburner 4.6.6.16555 beta 5 This OC setting pushed Speed Way to over 6000: https://www.3dmark.com/3dm/131777578 +500 MHz on GPU core crashed the card immediately when entering Speed Way.

New conditions: -25 mV undervolt on CPU (ALL cores plus cache through Intel XTU) +300 MHz GPU memory, +350 MHz GPU core through MSi Afterburner 4.6.6.16555 beta 5 Time Spy: 22 391 https://www.3dmark.com/3dm/131773151 Time Spy Extreme: 11 559 https://www.3dmark.com/3dm/131773338 Steel Nomad: 5 523 https://www.3dmark.com/3dm/131772990 Speed Way: 5 924 https://www.3dmark.com/3dm/131772897

3DMark results are in. Conditions: -20 mV undervolt on CPU (ALL cores plus cache through Intel XTU) +150 MHz GPU memory, +150 MHz GPU core through MSi Afterburner 4.6.6.16555 beta 5 nVidia 576.02 driver Time Spy: Time Spy Extreme: Steel Nomad: Speedway: Observations: GPU has no problems sustaining 170+ W, briefly peaking at 175.3 W during Steel Nomad as reported by MSi Afterburner/RivaTuner. GPU temperature never went above 69 degrees (nice). Just ask if you want to see a specific benchmark or custom conditions and such. Will test this eventually.

Don't think the TB5/DP Alt mode is on the iGPU. Running my Titan connected to my desktop display using USB-C and dGPU mode just fine. MSi Center warned that dGPU mode disables DP Alt mode - this is wrong.