Bernie_O Posted October 18, 2018 Share Posted October 18, 2018 cd /path/to/mountpoint/folder iozone -e -I -a -s 100M -r 1k -r 2k -r 4k -r 16k -r 128k -r 512k -r 1024k -r 16384k -i 0 -i 1 -i 2 Link to comment Share on other sites More sharing options...
wdtz Posted October 23, 2018 Share Posted October 23, 2018 I was also going to write about too high iozone numbers, but not as high as Bernie_O I see up to 1.6M with larger blocks, evo, evo plus, evo+, 32 & 64GB Interesting, numbers are about equal for 2 year old evo+ and new evo plus The sandisk 64 GB A1 is also about equal,,, BUT, but the 32 GB is shit (sandisk ultra A1) 1/6 the speed? less? I greatly regret buying 2 of them I also tested with the iozone parameter -s 500M, numbers were about the same really should only see 500 -600K, so something is wrong this is with a U1 or maybe U3 reader attached to usb3 Here 32 GB sandisk ultra A1 ,, ---slow, slow, slow 102400 4 1811 1745 7607 7610 7054 1384 102400 16 3327 4794 23309 23288 21946 4240 102400 128 17653 14829 90395 90565 67828 10891 102400 512 16577 15942 93699 93687 86111 13926 102400 1024 18142 15978 94203 94270 90049 14355 102400 16384 17585 15736 94533 94624 94274 20876 and comparison sandisk ultra A1 64GB ,, notice the "too high" numbers 102400 4 239202 281839 301375 296697 38806 38378 102400 16 562704 667356 708478 734044 144208 142922 102400 128 979088 1420801 1688792 1694830 790470 723220 102400 512 953029 1377012 1514590 1580078 1209057 1128835 102400 1024 975497 1362280 1555382 1569586 1377268 1218438 102400 16384 918773 1259275 1379822 1381468 1333371 1214032 iozone -e -I -a -s 100M -r 4k -r 16k -r 128k -r 512k -r 1024k -r 16384k -i 0 -i 1 -i 2 above 64GB with Bernie's dd time sh -c "dd if=/dev/zero of=testfile bs=100k count=50k && sync" 51200+0 records in 51200+0 records out 5242880000 bytes (5.2 GB, 4.9 GiB) copied, 153.192 s, 34.2 MB/s Link to comment Share on other sites More sharing options...
valant Posted November 2, 2018 Share Posted November 2, 2018 tkaiser, DDR50 is not the "slowest SD mode". It's a mandatory UHS-I mode, capable of delivering up to 50 MB/s. So it's fast, but it uses the lower frequency (50 MHz) giving power saving. SDR104 runs at 208 MHz. Quote 3.9 Ultra High Speed Phase I (UHS-I) Card UHS-I provides up to 104MB/sec performance on 4-bit SD bus with the single end driver interface. Card form factor is the same and existing connector can be used. 3.9.1 UHS-I Card Operation Modes • DS - Default Speed up to 25MHz 3.3V signaling • HS - High Speed up to 50MHz 3.3V signaling • SDR12 - SDR up to 25MHz 1.8V signaling • SDR25 - SDR up to 50MHz 1.8V signaling • SDR50 - SDR up to 100MHz 1.8V signaling • SDR104 - SDR up to 208MHz 1.8V signaling • DDR50 - DDR up to 50MHz 1.8V signaling Note: 1.8V signal timings are different from those of 3.3V In the above quote, pre UHS-I are DS and HS 3.3V modes. DS is "the slowest". DDR50 is a thing in fact - fast and power saving at the same time. Unlike SD104 it should be supported by every UHS-I card. This UHS-I "problem" is really bugging me. How messy it is on SBCs. I am happy to finally hear that Odoroid does support UHS-I modes. And Tinkerboard - I actually suspected that Asus could not just f&ck up such a good feature as UHS-I, which is already there in the SoC, and just waits for the proper implementation on the board side. But what about RockPro64? Do you know of whether the board lets the SoC's SD controller to do UHS-I? Even Rock64's rk3328 should be able of doing so. Heck, even Allwinner's A20 (!), R40, A64, and I am sure H2/H3/H5 too all claim UHS-I support... But what boards have implemented it properly? Link to comment Share on other sites More sharing options...
valant Posted November 4, 2018 Share Posted November 4, 2018 On 2/22/2018 at 2:07 PM, tkaiser said: Let's look how these things perform, this time on a new platform: RK3399 with an SD card interface that supports higher speed modes (requires kernel support and switching between 3.3V to 1.8V at the hardware layer). So results aren't comparable with the numbers we generated the last two years in this and other threads but that's not important any more... see at the bottom. does anybody here know if this has been implemented for RockPro64? I asked at the irc, but nobody seems to know. I suspect the answer is "no" (the same as with Rock64, that also should have a UHS-I capable SD host controller), but the certain answer would be better than guesses. Link to comment Share on other sites More sharing options...
kvic Posted November 5, 2018 Share Posted November 5, 2018 Renegade has the wiring on board for UHS-I. SDR104 works great for me. rock64 doesn't support UHS-I from what I read. Not sure about rock64pro. 1 Link to comment Share on other sites More sharing options...
Magnets Posted November 10, 2018 Share Posted November 10, 2018 Using performance gov on opi PC with Sandisk 16GB A1 on a usb card reader Command line used: iozone -e -I -a -s 100M -r 4k -r 16k -r 512k -r 1024k -r 16384k -i 0 -i 1 -i 2 Output is in kBytes/sec Time Resolution = 0.000001 seconds. Processor cache size set to 1024 kBytes. Processor cache line size set to 32 bytes. File stride size set to 17 * record size. random random bkwd record stride kB reclen write rewrite read reread read write read rewrite read fwrite frewrite fread freread 102400 4 3546 3573 5092 5086 4412 2948 102400 16 4588 8634 15272 15273 13885 7955 102400 512 21594 30277 36395 36542 35605 27344 102400 1024 28300 32080 36932 37035 36661 26242 102400 16384 30036 32860 37809 37933 37894 28773 iozone test complete. Purchased from aliexpress ($5) but the scratch & verify sticker says it's genuine My samsung evo 64gb went read-only recently (20 months old) and then would not boot. I copied to another card and ran fsck and all seems OK. I think the later evos were not as well made as the earlier batches. 1 Link to comment Share on other sites More sharing options...
Recommended Posts