Clevo 2019

Would it be possible to tdp mod the rtx to increase performance? I'm thinking it might be worth it for 4k.

 

See my answer to ssj92:

http://forum.notebookreview.com/threads/clevo-overclockers-lounge.788975/page-1692#post-10855876

 

Can you elaborate a bit on the power delivery? Clevo - Msi -Dell in the spoiler. I mean your meanings/thoughts. Thanks.

Spoiler: RTX design

 

: OOoo and again, another reason why the 20xx series will die quickly and the next improved 30xx series will be released very quickly (maybe 2019/2020? ).
Or some kind of upgraded version 21xx?

 

As I previously posted, increasing TGP (TDP for GPUs) won't accomplish anything unless you can scale the voltage higher and increase the core clock if the current power limit is not being exhausted. GPU-Z can show a PerfCap reason like VRel long before the power limit is exhausted. This is no different that setting power limits for the CPU higher in the BIOS or with XTU or ThrottleStop. If the voltage and core clock doesn't go up you can set the power limit as high as you want to and it will be ignored. It is only a limit, nothing more. Removing those limits is a good thing, but TGP only matters when you use up what you already have. To use more power you need more voltage to go higher on the core clock, and the ability to manipulate the voltage on 20-series GPUs (including desktop cards) is greatly hindered. The maximum voltage is capped very low. On my desktop 2080 Ti the max voltage is a mickey mouse 1.093V, and depending on the vBIOS you are using, getting it to actually stick there can be very difficult. Some of the vBIOSes I have test are incapable of holding the max voltage, and interestingly, those are the firmware with the highest TGP available.

 

The hardware is solid, but they are all 3 crippled by their firmware.

Clevo RTX 2080 uses 11 single FET phases
AW RTX 2080 uses 6 doubled FET phases
MSI uses 6 single + 2 doubled FET phases

Here a better picture of the MSI:

Spoiler

 

Nice to see another year with zero collaboration between vendors on standardization of MXM form factors... each laptop vendor marching to the beat of a different drum and doing their own thing, LOL. It is not as if having specs (and performance results) all over the board were not enough of an insult.

 

So if the "regular" 2080 draws 150W, how much does the Max-Q version require? As mentioned before, you see vendors selling it with a 180W PSU, which at first I thought being seriously underpowered. But, after reading your posts, it may even be enough?

For reference, see https://www.pcspecialist.co.uk/notebooks/defianceV-rtx-17/

 

2080 Max-Q is 80/90W.

How bad it will be depends if they add in solder on unlocked Intel BGA or not. And ewerything can be tuned with firmware.

Make it really bad is easy...

 

Just because it has a certain power limit doesn't mean it will actually use it. If it exhausts it and has power limit throttling, that certainly won't be a good thing. If it doesn't have power limit throttling and it is voltage limited it will be just as bad, or even worse, because it will limit the core clock speed. All by itself, the power limit doesn't really mean anything. It's just a number and it's only a bad number when the limit is actually a limiting factor. Same applies to voltage. Alienware saying their 2080 has 180W limit is probably just a marketing tactic to make it seem more impressive than it will actually be. If they don't give it as much or more voltage as everyone else selling the 2080 it may under-perform and the extra 30W will be meaningless.

 

nVidia .

 

1) VII performs at 2080 level, not at 1080 level.
2) Drivers since new dev team took over with Catalyst a few years ago have improved, with better DX performance and lowered overhead, stability, and in-game menu for monitoring and adjusting settings. Your opinion of AMD drivers is outdated, especially when combined with OverDriveN, which is a 3rd party replacement for Wattman. Visit the Guru3D Radeon forum for more info.

 

But, how well does it overclock? That piece is what always ends up in me scratch AMD off my list. I'd love it if I could actually buy an AMD product that overclocked better than an Intel or NVIDIA product and save a few bucks in the process. It has been many years since that was possible. I know that overclocking is often not a big deal to gamers, but it's almost the only thing I care about when it comes to computers. If I were a gamer, I probably wouldn't waste money on more expensive NVIDIA GPUs. It doesn't take a ton of horsepower for most games. That's why 1060/2060 and 1070/2070 are popular among gamers. They're "good enough" to get the job done without costing as much.

 

The actual news release is more specific. http://www.eurocom.com/ec/release(390)ec

Specifies two gpus, no SLI and an analogy about two snowblowers instead of a super snowblower. The notebookcheck advertisement is an odd riff.

Interesting use case for someone who needs dual rendering targets, (relative?) mobility, doesn't want to carry an epgu, does carry multiple screens and is terrible at planning. Maybe a traveling indie movie producer with an AI fetish who somehow has spare funds for dual MXM 2080s?

 

Won't know the details of overclocking the VII until it gets released. Suffice to say, OverDriveN works very well, allows adjusting of P States, and voltages. There's a nice Radeon OC community on Guru3D that'll be happy to help.

 

We should know pretty soon. I hope it will overclock like a banished with her hair on fire... Like 2300-2400 on core would be nice.

 

We need AMD back. That would make nVidia wake up, as it happened with Intel when Ryzen arrived.

 

They even caught up on single core performance. A ryzen 7 2700x overclocked to 4.1ghz has equal single core performance to a 7700k!
AMD need to teach nvidia a lesson just like they did with Intel!

 

War has its place for sure. To everything there is a season. Let's just hope it happens. I want to see raging rivers of red, green and blue blood gushing on the overclocking leaderboards and massive bickering, hate and contention spewing between rivals like the good old days. I do not really care what any of these parts run like stock.

If AMD would just open a can 55 gallon drum of overclocking whoop-ass in GPU land, NVIDIA would get pushed into a corner and have no choice but to knock off the nonsense with the room temperature thermal throttling, cancer vBIOS, and power and voltage limiting nonsense. And that, boys and girls, would be awesome for everyone no matter what color they prefer.

 

At a higher power consumption so the 7nm stuff will be interesting.

 

Nvidia will have that same choice, high power savings 7nm process or high performance 7nm process, unless they have access to a magical process that incorporates both?

It's going to take Nvidia a long time to clear the shelves of the $2B+ GTX 10-series inventory + the still new RTX inventory that piles up, before Nvidia can justify invest in releasing a new generation of 7nm GPU's.

Maybe Nvidia will be forced to drop prices finally, and clear out that inventory.

 

Where you see that?

 

7nm was shown at similar ipc per clock at lower power.

 

According to Nvidia, they've already ran out of 1080/1070s.

 

Whether that is so or not, there are still a lot of other sku's of Series-10 GPU's, hopefully not all of the least desirable models, or Nvidia is really in trouble if that Inventory isn't salable.

$2,000,000,000 worth of Inventory of GPU's, 120+ days of Inventory, not including what is being newly produced to release new RTX products. That is a lot of GPU's, it's gotta be something, right?

Maybe Nvidia should make a list of what they've got gathering dust on the shelves, list the specific sku's and numbers, and start a discount sale to clear Inventory - before the Inventory is too far past usefulness to bother selling.

Maybe if it's just "chips", and no AIB partners want to buy them, Nvidia can list them on various "craft's faire" sites and sell them to make jewelry and other crafts items?

27 Places to Sell Handmade Crafts Online
https://wahadventures.com/places-sell-handmade-crafts-online-free/

Remember those CPU Ties, and CPU bola's, maybe some belt-buckles and Hat pins? Something like those, and that way Nvidia can get the Inventory off the shelf and put to good use. Saving money on storage at least.

Somehow what Nvidia has been doing hasn't been working, so they've gotta start getting creative.

 

Do you a source for the 2 billion in inventory? I haven't seen it that high even after taking back all of those cards months ago.

 

looking at "right" chart? if we comparing 2700K and 7700K Clock to clock they're almost the same (but shintel clock higher )

 

Read my posts follow the links, all we have are estimates with a range higher and lower than $2B, based on data gathered, here is the latest estimate from SeekingAlpha, only $1.4B and I think it is conservative based on 4 months of inventory on hand, vs the 2-3 quarters quoted at various times from Nvidia - which all of a sudden vanished to " hey we caught up" recently, while Nvidia's stock was in free-fall, not long after claiming they were 2 quarters deep in unsold inventory and it was going to take that long to clear it.

Check out the link in my most recent post here, and then look for other posts and links in that thread and others:
http://forum.notebookreview.com/threads/nvidia-thread.806608/#post-10857721

Here's another "fib?" by Nvidia that some have sussed out recently, just how much % of sales was due to Crypto - Nvidia's number and the numbers resolved from reported data:

" Nvidia generated $1.95 billion in total revenue related to crypto/blockchain rather than its reported $602 million, according to RBC analyst Mitch Steves."

http://forum.notebookreview.com/threads/nvidia-thread.806608/page-174#post-10857310

I've made many other posts with interesting info about Nvidia's failings recently, check'em out.

 

So you really don't have a source for their inventory levels, just guestimates? I searched SeekingAlpha briefly and didn't see anything.

 

You didn't see that graph at the link in that Post? There are 5 pages in that article, look again, on page 2:
https://seekingalpha.com/article/4237475-nvidia-strong-market-disguising-downside-potential?page=2

I've been quoting estimates as I find them, and that's where the data is coming from.

If you want to call up Jensen and ask politely, maybe he will tell you? So far he isn't telling anyone else in enough detail to be exact.

Nvidia did say they came up Short $500M in estimated sales for last quarter, and admitted their new Turing GPU's aren't selling as well as expected, so if Nvidia didn't sell an expected $500M of Turing GPU's, they've got an additional $500M of new Turing Inventory to add to that $1.4B of old inventory (previous quarters inventory including Turing and Series 10) estimate, making $1.9B, close enough to round up to $2B for me.
http://forum.notebookreview.com/threads/nvidia-thread.806608/page-175#post-10857817

If you find better estimates, please do share them with all of us.

 

Thanks for the link, I did not see the second page. That guestimation was done November 18 2018, with an estimate of 124 days of inventory on hand (including Turing). I'm going to say that they've probably sold through some of that inventory in those 76 days...

 

Not likely with a $500M shortfall in sales for Q4, right?

Given Nvidia said it was going to take 2 quarters to catch up, then a couple of months later said they've almost caught up, right though Q4 when Nvidia said they had a $500M shortfall in actual sales vs estimated sales - on top of already downsizing their estimate for sales the month before - add to that Nvidia saying Turing GPU's weren't selling as well as they had hoped.

The "almost caught up" quote came right in the middle of an Nvidia stock price free-fall...

IDK, call me skeptical, but I think Nvidia still has $2B in inventory overall...including $250M in excess DDR memory. And, we haven't gotten a breakdown of what % of each sku are in inventory - and if as yet unreleased new products are counted in that inventory - which kept happening throughout that time period, adding new inventory on top of old.

With that $500M sales shortfall, I'd say new Nvidia products have been piling up on their shelves as unsold inventory faster than Nvidia is selling it, growing that inventory backlog, not reducing it.

Hey, let's see what Nvidia says on Feb 14th, maybe they'll pull it all together and give full disclosure by then?

Until then, check out my other posts on Nvidia over the last few months, there's lots of good info there.

 

If you saw or better say read @DaMafiaGamer’s post you could see he talked about oc’d 2700x Don’t read posts with the devils eyes

 

Sadly I don't have time to read the pages in that Nvidia post. I was curious what your source was. What I heard was that they were caught up/sold out in 1080s and 1070s, not with their whole product line. Anyway none of this really belongs in the Clevo forum, lol. I seem to be good at derailing posts. Thanks for providing the links.

 

P775TM-G Laptop Review - RTX 2080 + 9900K!

 

That 9900k is going to melt.

Maybe change the company name to MeltingBox? It would be a great tribute.

 

@Meaker@Sager the GPU temp goes up by 10C when the CPU is stressed simultaneously. That looks like a significant amount of heat bleed to me.

 

Heat bleed or lack of ability to tap into the other side of the cooling as it is now in use?

 

Are there any Sager models with new RTX cards but no HDD bay with the intent to minimize weight or maximize battery.

 

No, as generally most people go with an M.2 SSD + mech HDD for bulk storage.

 

Would it be possible to instal a der8auer Intel 9th gen. OC-Frame on the p870tm-R?
https://www.overclockers.co.uk/der8auer-intel-9th-gen-oc-frame-hs-008-dr.html

 

How would the GPU be able to tap into the CPU side cooling? Its shared heatpipe doesn’t connect to the other fin stack.

 

I am also worried why did he only get 23500 graphics points in 3DMark FireStrike, this is pretty low for an RTX 2080, my old factory overclocked gtx 1080 could get this.

Software issue ?
Overheating issue ?
Or bad power phase design ? (11 single phase fet)

 

Ops, I can say same for you my dear, sorry but We obviously speakin' of AMD 2700X cpu versus Intel 7xxx Gen. CPU
(so the posted graphical chart we can see Intel 7600K 152pts. vs. AMD 2700X 148 pts)

 

Most likely thermal because it is lower than other stock results of the same card in the P775TM ( ~25K) and P870TM ( ~27K) that were running at lower temps.

 

Not directly no, but it does go to a copper plate which then has 3 heatpipes on it going to that gin stack.

 

14nm vs 12nm is not a huge jump process wise and the shaders are similar.

Under purely shader bound tests the difference was never going to be dramatic.

 

Anyone got an answer for this?

Edit: I'm guessing you'd need a shim, and it might screw with the heatsink mounting.

 

A shim, so in effect you want to remove the IHS and put it back...