Lenovo ThinkStation P3 Tiny Gen 2 Performance
As noted earlier, the P3 Tiny can come with one of several Intel Core Ultra 2 Series CPUs. Our specific model uses the top-end Core Ultra 9 285, an 8 P core and 16 E core chip that is the pinnacle of the Arrow Lake platform, and allows the P3 Tiny to put up some very competitive numbers.
One of the parks of using desktop hardware here is that the Lenovo system gets access to more hardware resources overall. Whereas a mini-PC based on mobile hardware would top out at 6 P cores and 8 E cores (and 2 LPE cores), the desktop chips get access to another 8 CPU cores altogether. As a result, so long as the system can keep the more powerful and power-hungry desktop chip cool, it can easily outrun the more power-efficient mobile chips.
Though, as noted before, Lenovo does shoot itself in the foot a bit on the performance front by shipping the system with only a single CSO-DIMM, leaving the P3 Tiny short on memory bandwidth. Memory pricing has gone up a lot, but we went into The Difference between a Standard DIMM and a CUDIMM or CSODIMM using Crucial modules (when that was still a brand.)
Geekbench 6 Results
For our look at system performance, we will compare the P3 Tiny to a fellow Lenovo 1L tiny PC, the ThinkCentre M75q Tiny Gen 5. The AMD Ryzen 7 Pro 8700GE (Zen 4) system provides a useful benchmark for what these mini-PCs can do.
At a high level, the P3 Tiny and its Core Ultra 9 285 have little trouble trouncing the slightly older ThinkCentre system in both single-threaded and multi-threaded workloads. On average, the P3 Tiny is ahead by 33% in single-threaded workloads, and an even wider 59% in multi-threaded workloads.
Thankfully for Lenovo, there are no clear and obvious signs of the P3 Tiny being bottlenecked by its limited memory bandwidth. But I would be curious to see what another stick of RAM would do for some of these lower-performing multi-threaded tests.
Meanwhile, the GPU compute tests help to illustrate the significant performance difference between the CPU’s integrated GPU, and the discrete NVIDIA RTX A1000 card installed in the system. Even this low-end NVIDIA card is ahead by almost 300%.
Finally, the story is much the same for the Geekbench AI test. Using DirectML with the ONNX runtime, the A1000 is anywhere between 200% and 300% faster than the integrated GPU.
MLPerf 1.5
We ran MLPerf Client v1.5 on the system as well.
For AI workloads, the RTX A1000 comes in well ahead of the CPU and NPU in both latency and throughput.
Next, let us get to the power consumption and noise.
Can I populate all nvme slots at the same time? Could I fit double sided nvme? With PLP?
@Florian
Yes, all of the M.2 slots can be populated at the same time. As for double-sided drives, I’m afraid I don’t have a well-informed answer for you. Lenovo provides no official guidance in their manuals; they don’t explicitly list DS drives as supposed, but they don’t list them as unsupported, either. Thermal pads are pretty squishy, though, so I would be surprised if you couldn’t make it work.
It really is a shame they didn’t opt for a bigger GPU, it might have needed a custom heatsink but the b2000 would be a huge upgrade and should fit inside the systems power profile
I don’t disagree in regards to a more performant video card. But heatsink size aside, the RTX Pro 2000 Blackwell is not a drop-in replacement. NVIDIA specs it as a 70 Watt card, which is much higher TDP than the 50 Watt A1000. It’s only a difference of 20 Watts, but that’s a significant difference for such a small PC. At the moment NVIDIA does not have another 50 Watt video card; they’ve yet to release a newer 1000-tier card for the desktop market.
Oh I’m not pretending the 20 watts doesn’t matter although I’m sure that power supply can keep up with it, I would imagine that they would need to use the lower profile he’s thinking probably make it out of copper possibly out of vapor chamber it’s not it insurmountable wattage when laptops have cards in that power range although they do have the advantage of more integrated cooling solutions than what would have to be used here. I’m sure that the unfortunate reality is that it’s not a high enough volume product to invest the engineering resources required into anything more than the custom I/O bracket for what’s probably otherwise a standard 1000 class card.
Ugh that should say lower profile heatsink, autocorrect strikes again. (We need a better comment system that we can edit)
The computer actually supports up to 8 displays if you use daisychaining or a MST-hub on the onboard Displayport.
But the benefit of the nvidia card is of course only available through the four Mini displayports.
Gjetting bigger GPUs in this chassis is usually prohibited by the length of the card (or the heatsink arrangement on the card).
New on this model is the third nvme slot.
It’s a sturdy powerpack actually tested and certified for Maritime use.
The little computer that could
Why, in 2026, are these still being shipped with a single 1Gbps ethernet port? Sure, you could throw a faster NIC in the PCIe slot, but I’d rather use that for other purposes. Would love to see a 10GbE port, but even 2.5GbE would be a big improvement.
Really happy to see new TinyMiniMicro content … I’d also love to see links to new articles & videos posted on Mastodon, if you can spare the time. Lots of engagement from a strong homelabber crowd there. Great work as always!
I have been looking for something like this. I like it. It could be glued to the back of a 4K TV and used as a media machine. No extra box sitting around, out of sight.
Agh, 1GBe ethernet port…. why oh why is this a thing still! 20 years ago I can understand, but now? Integrating 10GBe onto motherboards will drive down the cost, and increase rate of adoption.
Also, why oh why is it DP1.4. DP2.1 is superior, offers higher resolution and display rate.
With 10GBe this could rival the Minisforum MS-01 and MS-A2, which are great machines but have drawbacks with remore management and heat dissipitation (and perhaps warranty replacement…), and are not 1U in height.
@Alan
The DP1.4 outputs are a function of the video card. The A1000 is an Ampere architecture product; Ampere doesn’t support DP2.x. That didn’t come until the Blackwell architecture (RTX 5000 series/RTX PRO Blackwell).
In fact, there simply isn’t an RTX PRO video card available that would fit into the P3 Tiny that offers DP 2.1. This is the best card available in the single-wide HHHL form factor.
Just an FYI – while it’s nothing even remotely approximating cost effective, there are a small handful of potential GPU upgrade options for these machines, most pertinently the RTX 2000E Ada Generation (note the “e”). Priced at around a frankly absurd $750 US, it performs roughly on par with (if not a few percentage points better than) the previous gen RTX A2000 in most cases.
It’s essentially, the same card as the RTX 2000 Ada Generation, except with a single-slot cooler, and BIOS Throttled to pull only 50w. Given that the 2000E costs the same as its beefier twin brother, it’s not a very compelling purchase…but it is a pretty substantial upgrade over the A1000. Worth mentioning that the gulf in performance between the 2000E ada and the 2000 ada is about the same as the gulf between the A1000 and the A2000.
@CantankerousRex
That’s a great find! Thank you for that.
When I was putting together this article, I was looking solely at NVIDIA’s official listings. I never thought to look at PNY’s listings.
As best as I can tell, the 2000E is a PNY-exclusive card. So I won’t hazard a guess as to what OEM availability is like. But if nothing else, that is a very interesting option as an end-user upgrade. Ada has a huge performance advantage at iso-power, to say the least.