From what I have seen at barefeats, write throttling (yes, I will use that word) of the SSD is a far bigger issue.
They were using the 256 GB version while the other machines were mostly 1 TB SSDs. Would be interesting to see how the performance scales with larger SSD models.
I wouldn’t call it throttling until more test results are provided. Barefeats said “The graph in AJA revealed major thermal throttling kicking in instantly, resulting in about half the average transfer speed expected.” That doesn’t make any sense as thermal throttling minimally implies some time frame for the chips to get hot, and it typically isn’t “instantly”. In a few seconds perhaps. And throttling of write speeds? But not read speeds? In an SSD? The T2 was throttled?
Meanwhile, Marco Arment showed 2700 MB/s write speeds on his testing using BlackMagic. His model was a 1 TB SSD model though.
From what I have seen at barefeats, write throttling (yes, I will use that word) of the SSD is a far bigger issue.
As far as I can tell, this isn't throttling based on thermals at all. This is the effect of smaller SSDs and a write evolution taking much longer than a read one, versus larger ones.
The same effect is seen in 128GB SATA SSDs versus 256, versus 1TB. The 1TB version of the same model drive just about always has faster writes than the 128 and 256.
Comments
I wouldn’t call it throttling until more test results are provided. Barefeats said “The graph in AJA revealed major thermal throttling kicking in instantly, resulting in about half the average transfer speed expected.” That doesn’t make any sense as thermal throttling minimally implies some time frame for the chips to get hot, and it typically isn’t “instantly”. In a few seconds perhaps. And throttling of write speeds? But not read speeds? In an SSD? The T2 was throttled?
Meanwhile, Marco Arment showed 2700 MB/s write speeds on his testing using BlackMagic. His model was a 1 TB SSD model though.