which earn me advertising fees or commissions.
As an Amazon Associate I earn from qualifying purchases.
2017 iMac Pro: Integrity Checker Verify
Related: 2013 Mac Pro, 2017 iMac 5K, 2017 iMac Pro, 4K and 5K display, computer display, CPU cores, data integrity, diglloydTools, iMac, iMac 5K, iMac Pro, IntegrityChecker, Mac Pro, Macs, SSD, storage, Thunderbolt
The data on this page and others was painstakingly obtained, made possible by the support of OWC / MacSales.com and B&H Photo. Please order through our links via these companies and/or subscribe—thank you.
See the MPG recommendations for iMac Pro, backup, peripherals. Not sure which Mac to get or how to configure it? Consult with MPG.
This test involves CPU cores as well as disk I/O speed.
IntegrityChecker(tm) v1.2.4 64-bit, diglloydTools 2.2.15
diglloydTools IntegrityChecker 'verify' verifies a large set of image files (SHA1 hash). IntegrityChecker is highly optimized for multi-scaling nearly perfectly out to 12 CPU cores, so I/O speed is a gating factor. At least has been the case; the native version of IntegrityChecker is impaired here and gave me fit in testing it. Apparently changes to macOS 10.13.2 and also 10.12.6 have mangled performance, possibly a result of fixes for the Meltdown and Spectre CPU bugs. Never in a decade has the native version been so impacted (no code change in the core algorithm for years and yet the performance droops with recent OS release).
Cores matter here, as shown by the scalability tests for SHA1.
Java code runs impressively fast on the iMac Pro, with throughput over 2.5 GB/sec for IntegrityChecker/Java version. Performance is severely degraded in the native version due to new performance bugs in macOS High Sierra, so the native version is not shown.
The 2017 iMac 5K CPU cores are rockin' fast and so it takes only 11% longer. But most of the sad difference is macOS not performing properly—new behaviors that did not exist before.
macOS High Sierra has some severe performance bugs—it is impossible to reliably optimize the Java version because macOS sometimes runs 1/3 the speed just minutes apart—with the same files—extremely frustrating new bugs. The times shown here are the best times achieved for each machine.
Results below with macOS 10.13.3 including the Apple Supplementary Update. See the results from more systems with macOS 10.12.2 further below.
Still, the SSD in the iMac Pro is not up to the challenge, with only about 2.7 GB/sec read speed. Too-slow disk I/O throttles the gains over the 2017 iMac 5K to only about 50% faster, even with the 10-core iMac Pro. The iMac Pro is not 'pro' in the sense that its SSD cannot deliver the goods for 8 or 10 CPU cores. That the SSD is the gating factor on the iMac Pro is proven by the scalability tests for SHA1, which show near-ideal scalability. It’s sad to see the iMac Pro hampered by an SSD less significantly slower for reads than the 2017 iMac 5K. Its anti-optimization of faster writes offers little or not benefit to the vast majority of users.
A vintage 2013 slow SSD as in the 2013 8-core 3.3 GHz Mac Pro hampers it (only about 1GB/sec) whereas the 6-core 2013 Mac Pro uses and SSD about 50% faster, albeit one only about half the speed of the 2017 iMac 5K.
The 2017 iMac 5K and iMac Pro all have very fast SSDs and thus can run at very high speed. The 2010 Mac Pro used a RAID-0 stripe of dual OWC Mercury AccelsiorProQ PCIe SSDs, so it too has a very fast SSD.
Recommended Items for iMac Pro and iMac 5K
See the recommendations page for details on why these items are recommended.