Lenovo ThinkPad T16 G1 AMD laptop review: Efficient and faster than Intel

[Redaktion]

The Lenovo ThinkPad T16 AMD is slightly rare among the 16-inch business laptops, since it also shows some ambitions toward being a slim workstation, with its Ryzen 6000 CPU. In this way, it is very similar to the ThinkPad P16s, and it is also not very different from it in other regards. Thanks to its WQXGA display and RDNA 2 GPU, it is also suited for simple multimedia tasks.

https://www.notebookcheck.net/Lenovo-ThinkPad-T16-G1-AMD-laptop-review-Efficient-and-faster-than-Intel.690088.0.html

[NikoB]

About the keyboard, not a word in the review except for the result. Why is it have 90%? Is it 1.5mm or 1.8mm?

[Benjamin Herzig]

Since this is a chassis we already reviewed before, you can find information on the keyboard here.

[kek]

It rubs me the wrong way when OEMs reuse motherboards from smaller models on bigger ones like they did here. It feels cheap as f*ck.

As for the CPU, I like the low temps, that looks really good.

[Plubious]

The Ryzen Pro Variants are clock locked but it's nice to start seeing laptops with just the Radeon 680M graphics and the Ryzen 6000 APUs support USB4/40Gbs but the OEMs can label that as USB4 and chose not to Implement the 40Gbs(TB3 Protocol support). Also if the Laptop supports USB4/40Gbs why the lack of external GPU testing from the Online Tech Press for Laptops that support that? I'd like to see just how powerful an external GPU could be utilized before the USB4/40Gbs bandwidth becomes a limiting factor.

I'll be looking for more affordable laptops to begin using just the Radeon 6000 APUs with only the Radeon 680M integrated graphics instead of always paring the 608M integrated graphics based APUs with some Discrete Mobile GPUs and negating the need for the 680M there. So more Laptops that ship with the Higher end 6000(rebranded to 7030 series now) RDNA2/12CU integrated graphics and if I need more Graphics horsepower then maybe use a USB4/40Gbs connected external GPU instead there when not on the go.

[NikoB]

Zen3+ does not have a USB40/TB4 SoC inside.
AMD's first chips with native support for USB40/TB4 (despite the lies of their marketers a year ago) are Zen4, and not all of them, but only the Phoenix 7x40 series with a "4nm" technical process. There are officially 2 built-in USB40/TB4 ports.

The Zen3+ SoC is connected to USB40 via a Parade 8830 interface chip. Just like the H version of Alder Lake, which also does not have built-in USB40 / TB4 ports.

Only U/P series have built-in TB4 ports. And that's why the i7 1260p is faster with an eGPU than the i9 12900HX. That has already been proven by tests.

The most disappointing thing for those who buy now is that TB4 will become morally obsolete in a year. Runs TB5 with pci-e 4.0 x4 (or pci-e 3.0 x8) and 120Gbps for DP ports. 2x60 and possibly 120Gb/s per monitor, but it's not clear because DP2.1 does not support more than 80Gbps according to the standard (UHBR20 lane mode).

Therefore, if there is something to sit on for another 1-1.5 years, it's better not to twitch and wait for TB5 and full-fledged DP2.0 (2.1) ports with 80Gb / s and full-fledged HDMI 2.1 in every laptop. This means that 8k monitors and real 4k @ 144-165Hz are flourishing on the market, i.e. the junk that is now sold on the market in lossless mode 4: 4: 4 144-165Hz with HDR on the "HDMI 2.1" port is not supported for a banal reason - this port is cut off to 24Gb / s in them and in the built-in / discreet.

So a radical change in video standards and eGPUs will begin in 2024 towards the end of it...

[Indiana]

Looks like this particular machine with the 2560 x 1600 display and 86 Wh battery would last 17 hours and 19 minutes, consuming 4.96 W on average.

Intel with the 1920 x 1200 display and the same battery is 14 hours and 6 minutes, consuming 6.1 W on average. How this would allow me to end up with the battery life of AMD + Low power display or Intel + 2560 x 1600 I have no clue. The data just isn't there.

[Miki]

As of March 2024 the long-term power limits have been decreased as follows:
Best power efficiency: 18 W -> 15 W
Balanced: 27 W -> 20 W
Best performance: 32 W -> 25 W

These limits are reached after about 5 minutes of benchmarking.

Looks a bit like planned obsolescence. At 32 W the CPU temperatures were kept below 80 °C so it's not thermal throttling, just an artificial limitation.

[NikoB]

This rather looks like an attempt to get rid of more specific warranty cases. First, the speed is advertised, leading to faster degradation of the power unit. Especially with Intel platforms (who would believe that the processor and motherboard of a laptop can withstand 170W on the processor even for a year?), and then quietly and sneakily they reduce consumption in order to prevent premature product failures and not be subject to an increased frequency of warranty cases. Although the artificial slowdown factor cannot be ruled out either.

Why are the criminal and vile moves of manufacturers in the computer hardware market painless for them?

Yes, because, unlike cars, where, as we all remember, vile manufacturers deliberately lowered emissions in tests (at the same time, engine torque and power suffered - but these were different tests!), and then, on the contrary, increased power and emissions.

When this was revealed in the auto market, what did these scoundrels do? Again, by order of the authorities, they forcibly reduced the engine power to a level that complied with emission requirements. As a result, they all immediately ran into class action lawsuits in which they were forced to pay huge sums to the affected consumers after updating the car's firmware.

It's just that there is complete chaos in the PC/laptop market with the connivance of vile officials and the simple idiocy of the public, stupid consumers who, having discovered this, do not try to force both the authorities and the courts to compensate them for the loss of product characteristics. Why don't they try?

Yes, because these very key characteristics of the product were NOT formalized in manufacturers' declarations or for purchase purposes by consumers (and this also depends on the consumer laws of the particular country!)

In order for consumers all over the world to buy exactly what they expect based on false reviews of the first batches or according to the characteristics that were in the first versions of the BIOS (or before the elimination of security "holes", after which performance suffers), we need to formalize all these requirements for hardware at the state level, obliging the same laptop manufacturers to declare all levels of laptop consumption in official company documents.

After such mandatory declarations of the same levels PL1/PL2, etc. any fraud with their changes in the future in the BIOS will be automatically punished by law, because this would be a violation of the contract with the consumer. And they will automatically be subject to class action lawsuits with huge losses, which will lead to the fact that they will have to think in advance about how to realistically solve the problem of long-term reliability and they will not be able to artificially slow down the hardware without compensating everyone who bought their laptops and other hardware.

[NikoB]

The same could be required to oblige laptop manufacturers to legally declare noise levels at certain loads. Etc. and so on.

All this is easy to formalize. But this is not being done. This means there is chaos in the market, which is convenient for producers, but not for consumers.

Maximum formalization of the extensive characteristics of goods always leads to a more civilized and more profitable market for consumers.

But this is always opposed by manufacturers and their corrupt political assistants.

Consider Oregon's just-passed right-to-repair law. It sounds beautiful, but in essence it is destroyed by vile details introduced by corrupt officials and politicians. Much like a similar good law in New York State was simply buried, diluted with many vile clauses that were introduced by corrupt authorities.

[Neenyah]

As of March 2024 the long-term power limits have been decreased as follows:
Best power efficiency: 18 W -> 15 W
Balanced: 27 W -> 20 W
Best performance: 32 W -> 25 W

These limits are reached after about 5 minutes of benchmarking.

Looks a bit like planned obsolescence. At 32 W the CPU temperatures were kept below 80 °C so it's not thermal throttling, just an artificial limitation.

Go to Device Manager > System Devices and find Lenovo PM Device. Uninstall it and reinstall version prior to 1.69.15.0:

https://pcsupport.lenovo.com/hr/en/products/laptops-and-netbooks/thinkpad-t-series-laptops/thinkpad-t16-gen-1-type-21ch-21cj/21ch/downloads/driver-list/component?name=Power%20Management&id=E1B533C3-16CA-4FBE-8BD8-FB5D7A57F431

That being said, it is not planned obsolescence, they just heavily adjusted (for the better) power curves so now it's basically flat line of CPU Package usage, same with thermals. Peak performance is slightly lower but average performance is slightly better so I like it this way on my X1 Carbon. Also literally no fan activity or noise whatsoever, I was curious if it was broken so I had to do a test in the BIOS to actually hear it.

This is the current graph (for me): https://imgur.com/Fyw6lKl
It starts at PL2 and then after 28 seconds if the load is still going it just flats at 23W for eternity, no drops (this was after 30 minutes). Same with other new(er) ThinkPads. No even fan noise as peak temps never exceed 81°C but they stay at exactly 80 (fan starts to spin at 81+).

It used to be like this: https://imgur.com/txgeJGD

So yeah, if you want the latter just uninstall 1.69.15.0 (and install the older version!) and you'll be fine.