How about the transcoding speed of the new m1 processor,
Forum rules
An Activity Log is required for support requests. Please read How-to get an activity log? for details on how and why this should be provided.
An Activity Log is required for support requests. Please read How-to get an activity log? for details on how and why this should be provided.
How about the transcoding speed of the new m1 processor,
How about the transcoding speed of the new m1 processor, has anyone tried it?
Re: How about the transcoding speed of the new m1 processor,
https://www.youtube.com/watch?v=gbB3mAtvyEw at seventh minute.
There isn't any good benchmark yet.
There isn't any good benchmark yet.
Re: How about the transcoding speed of the new m1 processor,
https://images.anandtech.com/graphs/gra ... 119353.png
In the Mac mini (with a ~20W TDP, the laptops will be lower), it's better than 1185G7, worse than the desktop chips for x264.
In the Mac mini (with a ~20W TDP, the laptops will be lower), it's better than 1185G7, worse than the desktop chips for x264.
Re: How about the transcoding speed of the new m1 processor,
You can pretty much ignore the Windows Result. It's completely invalid.
M1: 11.4fps
i7-1165G7 : 6.3fps avg
Two problems with the i7 result.
1. x264 and h264 decoders can easily saturate a quad core i7. Why is it only using 65% utilisation? Somethings not right there.
2. Why is the CPU clock so low at 1.85Ghz? It's base frequency should be no lower than 2.8Ghz all core.
Some preliminary results here: https://www.anandtech.com/show/16252/ma ... -m1-tested
Mixed bag across the board. Looking pretty good on the efficiency side.
M1: 11.4fps
i7-1165G7 : 6.3fps avg
Two problems with the i7 result.
1. x264 and h264 decoders can easily saturate a quad core i7. Why is it only using 65% utilisation? Somethings not right there.
2. Why is the CPU clock so low at 1.85Ghz? It's base frequency should be no lower than 2.8Ghz all core.
Some preliminary results here: https://www.anandtech.com/show/16252/ma ... -m1-tested
Mixed bag across the board. Looking pretty good on the efficiency side.
Re: How about the transcoding speed of the new m1 processor,
Transcoding with hardware acceleration is a little bit quicker than anything else I have tested. Which brings a small tear when the other things have more digits in their price. Transcoding in software only is remarkably quick, well for software transcoding that is.
In actual use the M1 mini is even more impressive; it is a phenomenal bit of kit.
My iMac Pro is in need of counselling and trauma therapy though. I'm still getting my head around the RAM usage - it has way more headroom with its 8GB than my 32GB iMac Pro when given the same set of demanding tasks. Time to unlearn a few things.
In actual use the M1 mini is even more impressive; it is a phenomenal bit of kit.
My iMac Pro is in need of counselling and trauma therapy though. I'm still getting my head around the RAM usage - it has way more headroom with its 8GB than my 32GB iMac Pro when given the same set of demanding tasks. Time to unlearn a few things.
Re: How about the transcoding speed of the new m1 processor,
I'd be careful with that thinking on RAM. While certain apps may use less for various reasons, data is still data. HandBrake will still use very similar amounts of memory and 8GB in some cases, simply isn't enough. macOS will page to disk. With faster SSDs these days, paging is far less noticeable but it does reduce SSD lifespan if you hit it too hard.
Re: How about the transcoding speed of the new m1 processor,
Posted this in another thread:
So here's my preliminary, unscientific results with my new M1 Mac Mini. Encoded an episode of Stargate Atlantis (from BluRay 1080p source with passthrough DTS audio) in h.265, medium, quality 24.
1. v1.3.3 using Rosetta > 1hr 17min
2. 1.4.0 beta 1 universal binary > 1h 04 min
So here's my preliminary, unscientific results with my new M1 Mac Mini. Encoded an episode of Stargate Atlantis (from BluRay 1080p source with passthrough DTS audio) in h.265, medium, quality 24.
1. v1.3.3 using Rosetta > 1hr 17min
2. 1.4.0 beta 1 universal binary > 1h 04 min
Re: How about the transcoding speed of the new m1 processor,
It would be interesting to see how it compares to the most recent non M1 mini. To see how much of a difference the new chip makes.jbellanca wrote: ↑Thu Nov 19, 2020 2:10 am Posted this in another thread:
So here's my preliminary, unscientific results with my new M1 Mac Mini. Encoded an episode of Stargate Atlantis (from BluRay 1080p source with passthrough DTS audio) in h.265, medium, quality 24.
1. v1.3.3 using Rosetta > 1hr 17min
2. 1.4.0 beta 1 universal binary > 1h 04 min
Re: How about the transcoding speed of the new m1 processor,
viewtopic.php?f=9&t=40722&p=193440#p193440
For anyone that has an M1 device and is willing to spare some free time, please read the linked guide and post your benchmarks here.
For anyone that has an M1 device and is willing to spare some free time, please read the linked guide and post your benchmarks here.
Re: How about the transcoding speed of the new m1 processor,
It can do CQ!!!
So, just ran a 5 minute chunk of an Avenger - Endgame 1080P remux using my normal H265 settings.
Baseline - iMac 9900k with Radeon 580x - 48GB RAM
M1 Mac - MacBook Pro with M1 - 8GB RAM
Software (H265 / RF20):
iMac - 6m22s
M1 Mac - 9m30s
VTB (bitrate):
iMac - 1m48s
M1 Mac - 34s
VTB (CQ)
iMac - None
M1 Mac - 33s
I'm going to run the normal Big Buck Bunny and Tears of Steel benchmarks next. Does anyone want have thoughts on VTB benchmarks for CQ or bitrate?
So, just ran a 5 minute chunk of an Avenger - Endgame 1080P remux using my normal H265 settings.
Baseline - iMac 9900k with Radeon 580x - 48GB RAM
M1 Mac - MacBook Pro with M1 - 8GB RAM
Software (H265 / RF20):
iMac - 6m22s
M1 Mac - 9m30s
VTB (bitrate):
iMac - 1m48s
M1 Mac - 34s
VTB (CQ)
iMac - None
M1 Mac - 33s
I'm going to run the normal Big Buck Bunny and Tears of Steel benchmarks next. Does anyone want have thoughts on VTB benchmarks for CQ or bitrate?
Re: How about the transcoding speed of the new m1 processor,
Sadly, Quick Sync based VideoToolBox is notably higher quality. That persists if the M1 is running in CQ or CBR.
Re: How about the transcoding speed of the new m1 processor,
ToS - HEVC - CBR 3000kbps:
MacBook Pro - Apple Silicon M1 VideoToolBox:
PSNR: 27.1627
SSIM: 0.9446
VMAF: 72.6932
i9 iMac - HD630 Quick Sync VideoToolBox:
PSNR: 27.0948
SSIM: 0.9401
VMAF: 73.7722
HD530 Quick Sync:
PSNR: 27.1137
SSIM: 0.9407
VMAF: 73.0349
AMD RX 5700XT VCE:
PSNR: 27.1451
SSIM: 0.9427
VMAF: 73.0871
Big Buck Bunny at 1080P - HEVC - CBR 1250kbps:
MacBook Pro - Apple Silicon M1 VideoToolBox:
PSNR: 25.3496
SSIM: 0.9263
VMAF: 56.1637
i9 iMac - HD 630 Quick Sync VideoToolBox:
PSNR: 25.1764
SSIM: 0.9189
VMAF: 58.7927
HD 530 Quick Sync:
PSNR: 25.2241
SSIM: 0.9163
VMAF: 57.4368
AMD RX 5700XT VCE:
PSNR: 25.2572
SSIM: 0.9289
VMAF: 60.9075
Interestingly, the M1 seems to be near the top if tested for PSNR and SSIM, but loses VMAF and to my perception.
I'm running some tests at other bitrates now.
MacBook Pro - Apple Silicon M1 VideoToolBox:
PSNR: 27.1627
SSIM: 0.9446
VMAF: 72.6932
i9 iMac - HD630 Quick Sync VideoToolBox:
PSNR: 27.0948
SSIM: 0.9401
VMAF: 73.7722
HD530 Quick Sync:
PSNR: 27.1137
SSIM: 0.9407
VMAF: 73.0349
AMD RX 5700XT VCE:
PSNR: 27.1451
SSIM: 0.9427
VMAF: 73.0871
Big Buck Bunny at 1080P - HEVC - CBR 1250kbps:
MacBook Pro - Apple Silicon M1 VideoToolBox:
PSNR: 25.3496
SSIM: 0.9263
VMAF: 56.1637
i9 iMac - HD 630 Quick Sync VideoToolBox:
PSNR: 25.1764
SSIM: 0.9189
VMAF: 58.7927
HD 530 Quick Sync:
PSNR: 25.2241
SSIM: 0.9163
VMAF: 57.4368
AMD RX 5700XT VCE:
PSNR: 25.2572
SSIM: 0.9289
VMAF: 60.9075
Interestingly, the M1 seems to be near the top if tested for PSNR and SSIM, but loses VMAF and to my perception.
I'm running some tests at other bitrates now.