M1's GPU performance is actually 1050~1060? Any sources?

[sunny5]

According to Geekbench, it is 1050~1060's level but the problem is it's just a benchmark, nothing more. Many PC users actually told me that M1's GPU is not even as close as 1050~1060 cause they couldn't find any actual software results and benchmarks proofs nothing especially with Geekbench. I do want to prove that M1 is quite good but so far, I couldn't find any actual test results with editing software. I know most games are not M1 native so it is very difficult to test it for gaming.

Are there any links or sources showing that M1 is actually perform as good as 1050~1060 with actual software? If not, then maybe we are misunderstood about its performance...

 

[Hexley]

If you're trying to settle an internet argument about it then best to avoid the argument all together.

Even when Apple Silicon gains better than RTX 3090 GPU performance at 100W power consumption for the whole Mac it wont make PC gamers switch to it as Apple's personal computer business model does not allow for the frictionless swapping out of parts.

Based on my readings more than 6 months ago the M1's GPU is the best overall iGPU of any platform until today. Why? Because Intel & AMD's market for iGPU are not willing to spend extra for it. For those applications their customers will buy a dGPU from AMD or Nvidia.

This does not answer your question but highlights some arguments are best not participated in.

 

[pshufd]

I have a Windows system with i7-10700 and GTX 1050 ti next to an M1 Mac mini. The 1050 runs hotter, supports three 4k monitors, uses more power. As far as performance goes, it's hard for me to tell. The CPU performance is more responsive on the M1 which the much higher single-core score implies. The i7-10700 can do more work though - at least the type that I do.

 

[sunny5]

I have a Windows system with i7-10700 and GTX 1050 ti next to an M1 Mac mini. The 1050 runs hotter, supports three 4k monitors, uses more power. As far as performance goes, it's hard for me to tell. The CPU performance is more responsive on the M1 which the much higher single-core score implies. The i7-10700 can do more work though - at least the type that I do.

Hmmm. Since you have both, it would be nice if you can test them with GPU centric software but cant think of which one to try... I only care about GPU performances.

 

[leman]

In benchmarks and games, it’s somewhere between a 1050 and 1050 Ti (desktop, not laptop). Check out data for GFXBench, 3DMark Wild Life Ultimate, Metro: Last Light, Warhamer Games, Baldurs Gates 3. In the Kater, which is a better optimized game, the performance is actually close to a desktop 1650.

In GPU-assisted tasks like video editing M1 is faster due to much lower data handoff latency.

 

[BigMcGuire]

Not exactly what you asked for but, I have an Intel MacBook (Intel i7-1068NG7) with an eGPU (RX 580). I can run my 4k Monitor on WoW/Starcraft II with high/ultra settings fairly comfortably.

My M1 MBP needs 1080p and high settings to run the same games acceptably on my 4k monitor.

The M1 is amazing - quiet, silent, barely gets warm, and can run off of my 26,800 mAh Anker battery for hours and hours and hours - something the Intel MacBook could not do .


The Intel MBP also has 32GB of ram and I use it for work. M1 MBP can't run my Windows applications in Parallels and while it's fast, it has periods of time where it sometimes takes forever to open applications. Booting - the Intel MBP is 2-3x as fast as well, strangely enough.

 

[pshufd]

Hmmm. Since you have both, it would be nice if you can test them with GPU centric software but cant think of which one to try... I only care about GPU performances.

I've only tested with Geekbench 5 and the GTX 1050 Ti scores a little bit more. I use some software that uses the GPU lightly but nothing that would stress either out. The reviews out there generally state that they are comparable and that's good enough for me. What I want more is the ability to drive more than two monitors.

 

[leman]

I've only tested with Geekbench 5 and the GTX 1050 Ti scores a little bit more.

I have a strong suspicion that Geekbench compute strongly favors GPUs with dedicated memory. They likely only measure the time it takes to run the kernel, but not the time it takes to transfer the data to and from the GPU memory.

 

[diamond.g]

I have a strong suspicion that Geekbench compute strongly favors GPUs with dedicated memory. They likely only measure the time it takes to run the kernel, but not the time it takes to transfer the data to and from the GPU memory.

Geekbench compute is only one side of the coin. You would also have to look at GXFBench or 3dmark for rasterization performance.

I mention it because AMD new GPU architecture isn't good with compute but is great with rasterization (the old stuff is good at compute but not so much with rasterization).

 

[JMacHack]

It’s a difficult thing to quantify because the architecture is very different from your standard box.

If a game isn’t available for both platforms, you’re out of luck. If there’s a port, there’s no guarantee that the port takes full advantage of the new architecture. Plus there’s the issue of Rosetta being a translation layer, which means any program running on it may vary in perf

The geekbench result is often repeated because it’s the closest thing to a fair comparison we have.

At the current venture, I’d say the best “real world” comparison would be to compare various video editing suites. Since DaVinci is now native it might be a good example.

 

[sunny5]

In benchmarks and games, it’s somewhere between a 1050 and 1050 Ti (desktop, not laptop). Check out data for GFXBench, 3DMark Wild Life Ultimate, Metro: Last Light, Warhamer Games, Baldurs Gates 3. In the Kater, which is a better optimized game, the performance is actually close to a desktop 1650.

In GPU-assisted tasks like video editing M1 is faster due to much lower data handoff latency.

Any links or sources? Otherwise, there is no prove to support that M1's GPU performance is between 1050~1060. Yes, it is great but I need info to prove it.

 

[leman]

Any links or sources? Otherwise, there is no prove to support that M1's GPU performance is between 1050~1060. Yes, it is great but I need info to prove it.

GFXBench and 3DMark Wild Life results are online on the respective websites. I would post links but I’m on mobile so it’s a bit of a hassle.

As to the gaming benchmarks, it’s a bit more complicated. Nobody ever did systematic benchmarks as far as I am aware. But YouTube has a bunch of videos for games I have mentioned.

 

[Argon_]

I saw a video on YouTube in which an M1 ran Minecraft compiled for ARM. That's something directly comparable. It can be a GPU demanding task with a shader pack.

 

[diamond.g]

GFXBench and 3DMark Wild Life results are online on the respective websites. I would post links but I’m on mobile so it’s a bit of a hassle.

As to the gaming benchmarks, it’s a bit more complicated. Nobody ever did systematic benchmarks as far as I am aware. But YouTube has a bunch of videos for games I have mentioned.

Shadow of the Tomb Raider has a benchmark mode. Just set all the settings the same and run it on two systems to compare.

I saw a video on YouTube in which an M1 ran Minecraft compiled for ARM. That's something directly comparable. It can be a GPU demanding task with a shader pack.

Is the Java version more GPU heavy than Bedrock?

 

[leman]

Shadow of the Tomb Raider has a benchmark mode. Just set all the settings the same and run it on two systems to compare.

Yeah, that would be a thing to do. Know anyone who owns the game and has the systems to run the benchmarks?

 

[Argon_]

Is the Java version more GPU heavy than Bedrock?

It is far less optimized, although nothing too demanding while stock. With shaders and packs it can be as GPU heavy as you want it to be.

Ex: Sildur's Vibrant Shaders

 

[diamond.g]

Yeah, that would be a thing to do. Know anyone who owns the game and has the systems to run the benchmarks?

Sadly I don’t have either GPUs. I have the game though. Lol.

 

[Homy]

The 2020 Mac Mini Unleashed: Putting Apple Silicon M1 To The Test

www.anandtech.com

 

[thedocbwarren]

Hard to compare as NVIDA will optimize their drivers for games. End of the day we have a underutilized GPU architecture on the M1 right now so you can't compare it. Base performance shows it's in line with the 1050ti. I have both and it does seem to be the case. And that is with a game using Rosetta on the mac side and Windows on the PC side. I will say the UI is so much faster and responsive using the M1 so Metal in macOS has a much better feel than the 1050ti in Linux or Windows.

The M1 beat the CPU performance of all my PCs and my MacBook Pro 16 i9. End of the day this is a superior platform.

 

[Homy]

Those pc users are just lazy so they can prove their point.
Metro Exodus at 1080p medium on M1 30-50 fps via Rosetta translation.
Metro Exodus at 1080p medium on GTX 1050 30-50 fps native Windows version.

 

[poorcody]

This video does real-world application comparisons:

 

[FreedomPenguin]

imo, the apple is perfect as is, it does pretty much every computing process people need for work/pleasure.
if you want gaming, I suggest a desktop or laptop windows replacement, I have a fully loaded m1 iMac and I bought a powerful laptop for windows just to play games.

shame, but it is what it is, and im not even mad.

 

[UBS28]

Who cares about the 1050? That is a super old GPU.

For the same money as the M1 MBP (16gb with 1TB SSD), you can by the AMD 14” Razer Blade which has higher multi-core performance and higher GPU performance with the RTX 3060 in a similar thin design.

 

[pshufd]

Who cares about the 1050? That is a super old GPU.

For the same money as the M1 MBP (16gb with 1TB SSD), you can by the AMD 14” Razer Blade which has higher multi-core performance and higher GPU performance with the RTX 3060 in a similar thin design.

What's the PL2 on the CPU and the TDP of the GPU?

I have an i7-10700 and the CPU has a TDP of 65 Watts but the PL2 is actually 228 Watts. The CPU is pulling 53 Watts at 40% CPU usage. Max since reboot is 92 Watts. I believe that the max CPU/GPU on the M1 is 20 Watts.

GPU makers aren't making low-end and mid-range GPUs anymore - they're just selling older cards because they're made on old processes where they can still get fab capacity which is why the 1050, 1060, 1080, 1650, 1660 GPUs are still relevant.

 

[leman]

Who cares about the 1050? That is a super old GPU.

For the same money as the M1 MBP (16gb with 1TB SSD), you can by the AMD 14” Razer Blade which has higher multi-core performance and higher GPU performance with the RTX 3060 in a similar thin design.

The Blade is a nice gaming laptop of course, but it’s not something I would buy for work. So again, it’s a weird point. Can one buy a better gaming laptop for less? Sure. Can one buy a similar productivity laptop that’s better at gaming? Barely…