Clevo 2020

The waiting is always the worst part

 

No, but the bridges they have burned in the past have made most companies go for this sort of testing where AMD is not funding it themselves.

 

I'm going nuts lol! I've been wanting A portable workstation for years so this is hardcore anticipation, Its worse than the night before going to the dealership to purchase my Passat TDI LOL!

 

I've got some engine parts on the way to get my engine rebuilt and it's the same feeling.

 

Mine arrived from RJTech this afternoon. Going to put it together soon.

 

I'm Soooo Jelly! My SSD, RAM,Carbonaut,Minus 8 have arrived, My 3600X should be here tomorrow if everything from amazon stays ahead of schedule and Kryonaut should arrive by Tuesday. How long did you have to wait for your laptop to arrive from the date it was processed? That's where I ordered mine too!

awesome, Could you check and see if the manual you have is more than 40 pages? I thought I may have had the grail but no sir its shorted like the rest.

 

I don't have download access to them, but they don't tend to be publicly released (officially) since they are copyrighted documents.

 

Makes sense, I'll just have to deal until the laptop arrives. I wish I wasn't so early to where there's hardly any in-depth information like what to expect in the bios etc, I know Der8auer has video and clevo also has them on the tube but I needs the nittiest of the gritty of the info.

 

Awesome, congratz! Let us know how it performs in 3Dmark FS/TS?

 

I am perplexed by some of the design choices made on this machine.

The NVMe drive looks to be running at 2x speeds, which I can only assume is done to allow the SATA drive to connect to the CPU instead of the chipset? I don't imagine there is an option to disable the SATA port in favor of 4x M.2 lanes? Or perhaps the lanes are split equally between the two M.2 slots instead of connecting one via the chipset?

Are all the USB ports connected via the chipset as well?

I'm really looking forward to hearing more of your impressions!

Edit: Nevermind, it looks like one of the PCIe ports is running in gen 2.0 and the other is running in gen 3.0. This makes way more sense! Still curious if the USB 3.1 ports are via chipset or CPU.

 

If you are confused you can always look up chipset diagrams for the platform. Typically the CPU just has PCI-E lanes coming from it.

 

Right, but the diagram does not answer my questions directly.

Assuming this is bog standard B450, there are 16x GPU lanes and 4x lanes split in one of a few configurations. Seeing a 970 EVO benchmark from one of the few reviews published clearly showed it was being limited to ~2GB/s. This can either be because:

• The 4x Gen3 CPU lanes are split in half between the two M.2 slots
  
• The 4x Gen3 CPU lanes are split between 2x NVME + up to 2x 6Gbps SATA
  
• The 4x Gen3 CPU lanes go directly to one M.2 slot and SATA + 4x Gen2 NVMe comes from the chipset. This is what I assume is happening on this board now, but don't know for certain.

When B450 was introduced, it advertised up to 4x USB 3.1 ports directly from the CPU. Ryzen 3000 CPUs updated this to 4x USB 3.1 Gen 2 from the CPU, though it may not be supported on this motherboard. I assume the USB 3.1 ports on this machine are from the chipset, but it is possible they could be supported by the CPU directly.

 

The PCIe Gen3 x4 is connected to the CPU, the PCIe Gen2 x4 is connected to the chipset. I am pretty sure the USB ports are all connected to the chipset, other than the back USB-C may be using the Nvidia USB-C controller.

 

Exactly, the chipset handles all the I/O other than the 20 other lanes from the CPU.

The CPU lanes are limited to PCI-E Gen 3 for routing and power reasons.

 

If you knew all i/o was from the chipset, why didn't you say it earlier? USB ports on this platform can come from the chipset or the CPU and this is the first I have seen it confirmed that all i/o comes from the chipset. Are you also confirming the USB-C port is also provided via the chipset and not the CPU or the Nvidia card, which win32asmguy left ambiguous?

 

The USB-C port does indeed connect to the B450 chipset. I found a USB-C to USB-A adapter and tested it out. It used to be that we could see schematics and parts list in the service manual, but that was not in the one that was provided for me on this model.

 

Have to wonder if AMD will ever get a Radeon or Radeon Pro mobile chip together that will work with their mobile CPUs?

Why rely on Nvidia?

 

They have the MSI Alpha 15. I am sure they will have more in the future.
There is also some Asus TUF with AMD card and an Acer Nitro with the 560X.

I know this thanks to my database.

 

Because if I help lead you to the answer you learn what to look for and how to find it so you can figure it out for any chipset. Plus that way you can check if I have missed something.

USB-C ports can have multiple inputs to them depending on their feature set, nothing stops the USB part being from a chipset and the displayport injected from the GPU for example.

 

Thanks for your help. Can always count on you to go digging for the answers.

 

Think it was an efficiency with power pr watt argument that came up.

 

Yes, in terms of performance per watt, support and drivers there really is not much competition at the moment.

I'd actually say the drivers are the biggest issue.

 

Now this is interesting, this is not the sort of stuff in the release info, now I checked through channels to the engineers and this was their response:

According to them, the Type-C is from the CPU, rest is chipset.

 

Schenker announces XMG Ultra 17 notebook with Core i9-10900K and RTX 2080 SUPER
https://videocardz.com/newz/schenke...tebook-with-core-i9-10900k-and-rtx-2080-super

 

Hope the 2080 super mxm 3.0 is compatible with the 751 but i presume the cooling will take a big hit lol.

 

The pcb layout remains overall the same from what I remember, so hopefully it only needs a bios update.

 

I am just scared about the 200W power draw . Wonder if the 330W PSU with me can keep up.

 

I've been running the P870TM1 w/ 9900K and KF at 5Ghz and powermodded + OC'd GTX1080 on a 330W power supply for about a year now. And I've ran the P775TM1 with the same setup for a while as well.

 

https://cdrdv2.intel.com/v1/dl/getContent/615211

There's something very wrong with the 8 and 6 core Loadline and Amps ratings. This is Skylake cores, not Pluto Lake...and I'm not going to go into my wall of text anymore.
I'll just say--reduce cores, you increase loadline and reduce ICCMAX. That's how it was on CFL.

 

I assume they know what they are doing with this.

 

I'll bet you right now $500 SOMEONE over there doesn't.
Want to take me up on it?

As @Papusan said, it's science.
If you have 10 cores, max 24 amps per core, you don't make 8 cores with max 30.5 amps per core, then 6 cores with max 24 amps per core again. Otherwise 9900k could have had 245A also, but at the 1.6 mOhm loadline, that would be:

1520mv - (1.6 * 245) =1.128v at 245 amps with 30 amps per core. No, you are not cooling 245 amps on ANY sort of cooling except sub-ambient.

Every previous generation of chips REDUCED the ICCMAX with each core reduction amount.
Go look at 9th gen yourself if you don't believe me. Do any of the 8th, 9th or even 7th gen KBL chips have 30 amps per core?

https://www.intel.com/content/www/u...core/8th-gen-core-family-datasheet-vol-1.html

Remember this is the company that said up to 4.8 ghz cock speed, before the graphic was pulled ...

 

The 8 core part should be 192A, right? Since they also keep the LLC at 1.1 mOhms for both 8/10 core 125w parts.
But, yeah 245A is a ton of current. Perhaps Intel will send out a Hailea HC-1000B with the 10900K press kits, lol.

 

Yes that's exactly what I told a contact.
The 8 core part should be 192 and the 6 core part should be 140. Notice 140 is right next to 138 of the 9600K and 8700K/8086K 6 core parts. And the 6 core loadline is also incorrect electrically. Unless these are not 14nm+++++++ skylake cores...

So why would 8 core be 245 amps?
Not only that, the loadline didn't change either. The loadline must change because these are guidelines for Intel's absolute maximum ratings.

Just do the math. The math is easy.
10900k:
(please ignore "offset capability" for SVID--that's for sub-zero users--note that Gigabyte disabled offset capability by default while I believe Asus enables it by default):

1520mv - (245 * 1.1) = 1.250v @ 245 amps)

10700k:
1520mv - (245 * 1.1) = 1.250v @ 245 amps. BUT THERE ARE TWO FEWER CORES! Current can NOT be the same.
So if we adjust for intel's mistake and lower ICCMAX ourselves:

1520mv - (193 * 1.1) =1.307v @ 193 amps. But this can't be right either, because 9900k was:
1520mv - (193 * 1.6)= 1.212v @ 193 amps. So you need to raise loadline when you lower ICCMAX.

to keep the same values as the 10 core: you get:
1520mv - (193 * 1.4)=1.249v (instead of 1.1 loadline I used 1.4 mOhm loadline).

Maths.

 

Looks like someone over at Intel is doing common core math.

 

Lol, everyone has "That" guy, but if they truly screw up an option like that then it will be slower than last gen.

 

Is the embargo for 10th gen May 13th?

 

Soon (tm)

 

I'm sorry but Duke Nukem Forever gave me trust issues with that phrase.

 

Basically it can mean any length of time lol.

 

Well for DN:F that turned out to be 15 years. At least der8bauer has already done ln2 testing and can share that when embargo lifts.

 

Taken from 51M R2 thread found by Falkentyne, some 3DMark physics numbers with the 10gen chips.

 

Hardware existing is always a good sign

 

DN:F is kinda a funny comparison to the 10th gen, both are archaic even before launch.

 

Maybe if DNF was a bit more fun at least, you'll still get a good experience with these CPUs in the likes of gaming and everyday tasks.

 

If nothing else, you might be able to snag a cheap 9900KS after this launch.

 

Still the top chip for the platform so maybe but maybe not.

 

If current pricing is any indicator, that would be a no. 9900k's and kf's (outside of binned 5ghz+ chips) can be had extremely inexpensive, but the ks variant are still netting 700-1000/ea.

Limited run and still the best you can get without slogging through a bunch of 9900k/kf binning.

 

I've seen a few pop up for sale on local hw forums as people are worried that the resale value might drop when 10th gen launches. Not that they were cheaper than the current 9900K

 

Yeah I think it will trend down a little but not much. Not enough to tempt me from my 8086k.

 

Well the vapor chamber direct die mod has thermal ceiling for 9900K at 5.2Ghz in the P870, so I'd imagine you could daily something like 5.3/5.4Ghz with the 8086K.