128 bit Encrypted disk image

[Loa]

Hello,

I've been using a couple of 128bit encrypted disk images, and I do love their level of security.

What I don't like is their slowness. Copying to or from these images is painfully slow. A couple of gigs will take 5-6 hours.

thing is, I don't see any CPU activity, and the disk activity is in the KB/s range.

What is slowing things down that much?

Loa

 

[webs2slow4me]

Same issue. Wondering if you ever figured it out.

 

[h9826790]

May be this is the reason

 

[bookemdano]

AES-NI is magical. Highly recommend a CPU upgrade.

 

[Loa]

Yeah, I do have a w3580... Maybe I should have spent a bit more to get the w3690 when I bought it to upgrade my mac!

 

[AidenShaw]

Yeah, I do have a w3580... Maybe I should have spent a bit more to get the w3690 when I bought it to upgrade my mac!

According to that chart - the w3580 can do 145 GB/sec. Not sure that would explain your KB/sec.

On the other hand, the chart says that the w3690 can do 1.8 TB/sec - which is far more than the memory system can support.

Those numbers might be garbage.

 

[handheldgames]

comparing single core AES benchmarks in Geekbench 4...

w3580 115.8 MB/sec
x5677 1.47 GB/sec $30
x5690 1.47 GB/sec $200

All you need is a $30 x5677

 

[Loa]

Well, the numbers are all over the place. Is the w3580's AES score 144GB/s or 0.115GB/s?

Can Geekbench 3 and 4 give such ridiculously different test results?

 

[killawat]

Well, the numbers are all over the place. Is the w3580's AES score 144GB/s or 0.115GB/s?

Can Geekbench 3 and 4 give such ridiculously different test results?

The presentation is a bit strange but proportionally, the results are the same (Geekbench 3 = 1800000/144950 = ~12.41, Geekbench 4 = 1.47/.115 = ~12.7)
it is odd to express AES capabilities in terms of MB, I think time given a particular operation would be more apt comparison. The point is, hw acceleration offered by w3690 (and other AES-NI accelerated workloads) are much faster AES operations.

 

[handheldgames]

Well, the numbers are all over the place. Is the w3580's AES score 144GB/s or 0.115GB/s?

Can Geekbench 3 and 4 give such ridiculously different test results?

I took the numbers from the single core AES scores on Geekbench 4 benchmarks from the noted processors.

Re 3 vs 4....
Geekbench 3 and Geekbench 4 scores are not comparable. This is because Geekbench 4 not only uses a different baseline to calculate it's scores, but it also runs new and updated CPU and GPU workloads. You can read more about these workloads here, and here.
[doublepost=1535075775][/doublepost]

The presentation is a bit strange but proportionally, the results are the same (Geekbench 3 = 1800000/144950 = ~12.41, Geekbench 4 = 1.47/.115 = ~12.7)
it is odd to express AES capabilities in terms of MB, I think time given a particular operation would be more apt comparison. The point is, hw acceleration offered by w3690 (and other AES-NI accelerated workloads) are much faster AES operations.

AES Single core @1.60 GB/sec with the $90 w3690 is slightly faster than the $30 x5677, and the $100 x5690 @ 1.47 GB/sec