The company I work for has always used 15" Macbook Pros, because they were always the best combination of power and portability for our needs. We use these machines to host and run different Virtual Machines. The choice of Virtualisation software is open - some use Fusion, others Parallels, and others VirtualBox.
A typical requirement might be to build a couple of VMs, such as two Windows Servers, install 3rd party software and set up a demo, then save all this set up, turn up at a customer site, fire everything up and run the demo on site (many places have zero internet for cloud based demos).
We've had a collection of Retina MBPs, all with 16Gb RAM and 750Gb/1TB SSD. The 2012 models were a bit flaky, but the 2013-2015 models have been rock solid.
Now, the problem:
Two weeks ago, we bought two 15" brand new 2017 models, both with 16Gb RAM, Radeon chips, one with 500Gb SSD and the other with 1TB SSD.
On to each machine we have installed the latest VMWare Fusion, and set up one Windows Server 2012 VM on each - 2 cores, 6Gb RAM. Oh dear. The VMs hang and/or crash intermittently, and on occasion may bring the whole MBP to a dead stop.
The console shows many instances of the dreaded Kernel...gpuRestart problem. Its always the Intel Graphics GPU that fails. The failing program listed was either 'vmware-esx' or 'WindowServer'.
We had this previously on a 2012 rMBP which resulted in a board replacement, then an entire machine replacement, which fixed the error.
After much troubleshooting, one MBP was returned to Apple and replaced with a brand new one. Same problem.
Blame VMWare? Well here's the thing: We got the same symptoms running the same VM, on the same machines, but using VirtualBox instead.
If you seach for 'gpuRestart' you'll find this problem on and off for years. My suspicion is it may be a problem with the switching between GPUs. We wanted to try 'gfxcardstatus' to force use of one GPU, however it seems to have stopped development?
Unfortunately then, we may be forced to return these 2017 models and pick up two of the previous generation machines.
Bah.
A typical requirement might be to build a couple of VMs, such as two Windows Servers, install 3rd party software and set up a demo, then save all this set up, turn up at a customer site, fire everything up and run the demo on site (many places have zero internet for cloud based demos).
We've had a collection of Retina MBPs, all with 16Gb RAM and 750Gb/1TB SSD. The 2012 models were a bit flaky, but the 2013-2015 models have been rock solid.
Now, the problem:
Two weeks ago, we bought two 15" brand new 2017 models, both with 16Gb RAM, Radeon chips, one with 500Gb SSD and the other with 1TB SSD.
On to each machine we have installed the latest VMWare Fusion, and set up one Windows Server 2012 VM on each - 2 cores, 6Gb RAM. Oh dear. The VMs hang and/or crash intermittently, and on occasion may bring the whole MBP to a dead stop.
The console shows many instances of the dreaded Kernel...gpuRestart problem. Its always the Intel Graphics GPU that fails. The failing program listed was either 'vmware-esx' or 'WindowServer'.
We had this previously on a 2012 rMBP which resulted in a board replacement, then an entire machine replacement, which fixed the error.
After much troubleshooting, one MBP was returned to Apple and replaced with a brand new one. Same problem.
Blame VMWare? Well here's the thing: We got the same symptoms running the same VM, on the same machines, but using VirtualBox instead.
If you seach for 'gpuRestart' you'll find this problem on and off for years. My suspicion is it may be a problem with the switching between GPUs. We wanted to try 'gfxcardstatus' to force use of one GPU, however it seems to have stopped development?
Unfortunately then, we may be forced to return these 2017 models and pick up two of the previous generation machines.
Bah.