Hands On - iMac Pro 10 Core/64GB Ram/Vega 64

[shadow82x]

Decided to post some initial thoughts about my new iMac.

• Performance: Performance seems solid so far with geekbench showing promising results. (https://browser.geekbench.com/v4/cpu/5803907) Respectable single and multi-core speeds.
• Graphics: I need to spend more time testing but so far the Pro Vega 64 screams. Fired up some games such as Borderlands 2, Overwatch, PubG and it consistently had above 60 FPS. I will be testing in 5k resolution shortly. This machine will definitely be nice if you want to do some casual gaming with high frame rates.
• Noise: Some minor coil wine is present when writing files, nothing major or really noticeable. Other than that it's pretty darn quiet, quieter than my 2017 MacBook Pro 15" in all scenarios.
• Backlight Bleed: There is some minor backlight bleed on the bottom right and top left border of screens, noticeable on black screens. Annoying, but not a deal breaker. That's the risk with IPS LCD panelsI suppose as my MacBook has similar bleed. (Video shown doesn't do justice as that was a black screen and it appears to have a blue tint.)
• Size Profile (Irrelevant): I need a bigger desk because this thing is huge but the dimensions are same as 27" iMac.

So far, I'm pretty happy with the thing. My only real complaint is the very minor backlight bleed, but overall I'm satisfied. Feel free to ask me anything. I'll be performing additional benchmarks shortly.

 

[FredT2]

Run Handbrake and see what Intel Power Gadget says about temperature and processor speed.

 

[Mac32]

Congratulations with your iMac Pro!

Could you check if you are able to use the HDR setting in newer games (like Battlefield 1, Mass Effect Andromeda etc.), and if the iMac Pro screen supports this function? That would be greatly appreciated, I've been wondering about this. (You might need Windows 10/Bootcamp for this though.)

Looking forward to seeing some gaming benchmarks. The CPU's base clock is 3ghz, but it has a whopping 1.5ghz of turbo boost. I curious to how that works out in gaming performance. Enjoy!

 

[shadow82x]

Congratulations with your iMac Pro!

Could you check if you are able to use the HDR setting in newer games (like Battlefield 1), and if the iMac Pro screen supports this function? That would be greatly appreciated, I've been wondering about this. (You might need Windows 10/Bootcamp for this though.)

Looking forward to seeing some gaming benchmarks. The CPU's base clock is 3ghz, but it has a whopping 1.5ghz of turbo boost. I curious to how that works out in gaming performance. Enjoy!

Fired up some Pubg in Bootcamp with Intel Power Gadget. I'm happy to report that the game has maintained a constant boost of 3.7-4.2 GHz for my 30 minutes of game play with FPS ranging from 40-100. (~60 average at 5K resolution!) I don't own BF1 unfortunately, at least not yet. I'll be testing more soon.

 

[Mac32]

Fired up some Pubg in Bootcamp with Intel Power Gadget. I'm happy to report that the game has maintained a constant boost of 3.7-4.2 GHz for my 30 minutes of game play with FPS ranging from 40-100. (~60 average at 5K resolution!) I don't own BF1 unfortunately, at least not yet. I'll be testing more soon.

Wow, that's impressive, and speaks highly of the new thermal system in the iMac Pro. If you could also test the iMac Pro for HDR support in games (and HDR video content, for instance seen in Netflix), that would be fantastic. There are various new games that have the HDR option btw., maybe you already own one..

 

[dimensional]

Did you install any Vega drivers beyond what BootCamp installed? I was getting terrible performance in Overwatch (only Windows game I tried so far) while OS X games were getting about double the FPS I got on a iMac 2016. I have the Vegas 64.

 

[tillsbury]

Any chance of some bootcamp performance stats? Yours is the first mention I've seen of anyone running Windows on the iMac Pro...

 

[shadow82x]

No change in drivers. AMD says the latest are installed.
[doublepost=1514355466][/doublepost]

Any chance of some bootcamp performance stats? Yours is the first mention I've seen of anyone running Windows on the iMac Pro...

Any kind of software to test such?

 

[Trebuin]

Decided to post some initial thoughts about my new iMac.

• Performance: Performance seems solid so far with geekbench showing promising results. (https://browser.geekbench.com/v4/cpu/5803907) Respectable single and multi-core speeds.
• Graphics: I need to spend more time testing but so far the Pro Vega 64 screams. Fired up some games such as Borderlands 2, Overwatch, PubG and it consistently had above 60 FPS. I will be testing in 5k resolution shortly. This machine will definitely be nice if you want to do some casual gaming with high frame rates.
• Noise: Some minor coil wine is present when writing files, nothing major or really noticeable. Other than that it's pretty darn quiet, quieter than my 2017 MacBook Pro 15" in all scenarios.
• Backlight Bleed: There is some minor backlight bleed on the bottom right and top left border of screens, noticeable on black screens. Annoying, but not a deal breaker. That's the risk with IPS LCD panelsI suppose as my MacBook has similar bleed. (Video shown doesn't do justice as that was a black screen and it appears to have a blue tint.)
• Size Profile (Irrelevant): I need a bigger desk because this thing is huge but the dimensions are same as 27" iMac.

So far, I'm pretty happy with the thing. My only real complaint is the very minor backlight bleed, but overall I'm satisfied. Feel free to ask me anything. I'll be performing additional benchmarks shortly.

Could you run a yes test with Intel power gadget running to capture the temps & CPU speed? I did it on the 8 core & was surprised it did not down clock.

Open Terminal
insert the following command 20 times for a 10 core:
yes > /dev/null &
to stop when you've seen enough, insert the following command:
killall yes

 

[lorddevil]

Hey there, I have the same configuration except that I went for 32GB and 56 Vega.

I just ran your yes test with power gadget and I can say that it keeps around 3.2-3.3 for all 10 cores under full load - even the fan is much less noisy than it was on my old imac 4.2ghz.

But what is very interesting, the 10x core should turbo boost to 4.5 - at least form its specs. If I do only 1 or 2 "yes" to occupy exactly 1 logic processor - the Frequency reaches max 4.2 - the same as the 8x core. This seems like it got wrong advertised - and turbo boost is only 4.2 not 4.5?

I tried to do single task operations too and the gadget is never reporting anything higher than ~4.2ghz?

 

[Mac32]

No change in drivers. AMD says the latest are installed.
[doublepost=1514355466][/doublepost]
Any kind of software to test such?

You should be aware of this website: https://www.bootcampdrivers.com
However, iMac Pro isn't yet supported by these unofficial Bootcamp drivers, but seems like it's in the works.

 

[thefriendshipmachine]

1. according to istats or temperature gauge what is the CPU at after running the "yes" terminal command after 20 seconds, 1 minute and 3 minutes?

2. do you see any signs of early image retention as per this test https://marco.org/rmbp-irtest.html

 

[bplein]

But what is very interesting, the 10x core should turbo boost to 4.5 - at least form its specs. If I do only 1 or 2 "yes" to occupy exactly 1 logic processor - the Frequency reaches max 4.2 - the same as the 8x core. This seems like it got wrong advertised - and turbo boost is only 4.2 not 4.5?

I tried to do single task operations too and the gadget is never reporting anything higher than ~4.2ghz?

The "yes" is a simple test that doesn't use the complete CPU, it just keeps it busy. You should not assume anything about the TurboBoost capability based on a simple test like this.

 

[lorddevil]

You are probably right bplein.

Still I have just taken a look at your picture with your 8 core configuration. When comparing my Intel Power Gadget output with 20 Yes threads running and yours its totally different.

Here you see its switching constantly in a zig zag between 3.0 - 3.3 - 3.5 up down up down... while your cpu keeps the 3.9 absolutely steady and without any switching - I guess thats because it is reaching 160 Watt and needs to throttle.

But still it keeps around 10x3.3 which is great for this small form factor
[doublepost=1514408560][/doublepost]Oh and I just tested it with cinebench using only 1 core - its the same capping at 4.2ghz so in my eyes the 4.5ghz seems to be a wrong spec (!?) - which would be really evil because they advertise the higher priced processors with the gimmick of hitting 4.5 turbo instead of 4.2....

 

[FredT2]

You are probably right bplein.

Still I have just taken a look at your picture with your 8 core configuration. When comparing my Intel Power Gadget output with 20 Yes threads running and yours its totally different.

Here you see its switching constantly in a zig zag between 3.0 - 3.3 - 3.5 up down up down... while your cpu keeps the 3.9 absolutely steady and without any switching - I guess thats because it is reaching 160 Watt and needs to throttle.

But still it keeps around 10x3.3 which is great for this small form factor
[doublepost=1514408560][/doublepost]Oh and I just tested it with cinebench using only 1 core - its the same capping at 4.2ghz so in my eyes the 4.5ghz seems to be a wrong spec (!?) - which would be really evil because they advertise the higher priced processors with the gimmick of hitting 4.5 turbo instead of 4.2....

See what happens with Handbrake (I know, broken record). It causes the processors to run more slowly on my 2014 i7 than does running Yes.

 

[Trebuin]

You are probably right bplein.

Still I have just taken a look at your picture with your 8 core configuration. When comparing my Intel Power Gadget output with 20 Yes threads running and yours its totally different.

Here you see its switching constantly in a zig zag between 3.0 - 3.3 - 3.5 up down up down... while your cpu keeps the 3.9 absolutely steady and without any switching - I guess thats because it is reaching 160 Watt and needs to throttle.

But still it keeps around 10x3.3 which is great for this small form factor
[doublepost=1514408560][/doublepost]Oh and I just tested it with cinebench using only 1 core - its the same capping at 4.2ghz so in my eyes the 4.5ghz seems to be a wrong spec (!?) - which would be really evil because they advertise the higher priced processors with the gimmick of hitting 4.5 turbo instead of 4.2....

The thermal limit is probably around 93-94C, which is probably where it should be. Once we get a good fan control, I suspect we can ramp up the fan, drop the temps & see higher...hopefully.

 

[ivyharriet]

If you could post some more video game real-life tests I would really, really, really appreciate it! Ideally graphics settings screenshot and a few seconds of video or just an account of the framerate. I'm considering buying the Vega 64 iMac Pro for pro and gaming purposes but am worried it'll be underpowered for modern games at a high resolution.

 

[Mac32]

If you could post some more video game real-life tests I would really, really, really appreciate it! Ideally graphics settings screenshot and a few seconds of video or just an account of the framerate. I'm considering buying the Vega 64 iMac Pro for pro and gaming purposes but am worried it'll be underpowered for modern games at a high resolution.

I don't have the iMac Pro myself, but I've noticed that the iMac Pro bootcamp drivers atm (esp. for gaming) are lacking, so any benchmarks this early are going to be hit and miss. Your best bet is www.bootcampdrivers.com, which releases more up-to-date unofficial drivers, since AMD/Apple only updates their GPU drivers for Bootcamp every 1-2 years. Bootcampdrivers.com has not included iMac Pro in their drivers yet.

 

[lorddevil]

When I saw the differences between vega 56 and 64 for me it was clear that I won't go for the 64.The gain of roughly 10% average doesn't make sense to me for the price. If someone really want graphics power for games etc. he would not buy an imac pro in the first places. The performance of the VEGA is too bad in comparision to nvidia graphic cards. So if I really want power in the future I would get a strong eGPU instead of the VEGA anyway - thats why for me vega 56 was a clear choise.

 

[Double-Slit]

When I saw the differences between vega 56 and 64 for me it was clear that I won't go for the 64.The gain of roughly 10% average doesn't make sense to me for the price. If someone really want graphics power for games etc. he would not buy an imac pro in the first places. The performance of the VEGA is too bad in comparision to nvidia graphic cards. So if I really want power in the future I would get a strong eGPU instead of the VEGA anyway - thats why for me vega 56 was a clear choise.

So the iMac Pro (8 or 10 core / vega 64) is not suitable for playing VR games?

 

[SoyCapitanSoyCapitan]

All in one workstations aren't suited long renders due to their slim form factor but the great thing is the eGPU support will allow you to connect a GPU render box. We need an eGPU box with 2-3 PCIE slots for best results. Even the TB3 connection is sufficient bandwidth.

 

[lorddevil]

So the iMac Pro (8 or 10 core / vega 64) is not suitable for playing VR games?

Depends on how you see suitable - I would never buy a mac for playing a game - because any windows pc with a high end video graphics card will outperform by far with half the price .

But sure you can play VR games with it - still its a workstation and thus better suits other areas

 

[Double-Slit]

Depends on how you see suitable - I would never buy a mac for playing a game - because any windows pc with a high end video graphics card will outperform by far with half the price .

But sure you can play VR games with it - still its a workstation and thus better suits other areas

Blimey, again no reason to buy an iMac Pro ;-)

 

[shadow82x]

Did you install any Vega drivers beyond what BootCamp installed? I was getting terrible performance in Overwatch (only Windows game I tried so far) while OS X games were getting about double the FPS I got on a iMac 2016. I have the Vegas 64.

I stand corrected, spent some more time in Overwatch and can't get above 10fps no matter the setting. Odd, considering the GPU/CPU can easily handle the game. Must be a driver issue.

 

[Trebuin]

I stand corrected, spent some more time in Overwatch and can't get above 10fps no matter the setting. Odd, considering the GPU/CPU can easily handle the game. Must be a driver issue.

If I had overwatch, I’d look at it. This reminds me of what happened with Vista; I think it was NVIDIA chose not to develop drivers until after the release. Took about 6 months to get caught up & then games started working. Could be the same issue here.