Jump to content

dreamcat4

Members
  • Content Count

    116
  • Joined

  • Last visited

About dreamcat4

  • Rank
    InsanelyMac Geek
  1. Well you can ask gigabyte if they will replace the BIOS for you for a surcharge. However you will also have to pay shipping for sending your mobo to them. And back again (both ways). Unless you happen to live close enough to the Gigabyte RMA centre in your country / region whereby you could go there to drop it off in person. Assuming they are OK with doing it for you. Another option would be to search for 'BIOS CHIP GA-Z170MX-GAMING 5' on ebay. Those people are based in the netherlands. Its not that expensive and they include shipping. However you will also need to replace one of the 2 BIOS chips on your motherboard. Its not too hard but you need to have good-enough soldering skills. Another thing you can buy yourself is a SOIC-8 flash clip. Those things can be wired into an rPi or similar device, to create a hardware flasher. And then use that to reflash your existing BIOS chips in-situ without having to desolder them. In fact if you bought both products you could read what was on product A and then flash it onto product B without doing any soldering... and then make also a backup of it for later use with your hardware flasher should you ever end up bricking the BIOS again. I would also research on why situations occur when both the BIOS chips get overwritten with a bad image / setting. It shouldnt really ever happen. Even if you have downgraded as you say (and only downgraded on one of them. Maybe what normally happens is that on the 2nd successful boot of a new BIOS image, if its still good then it will automatically overwrite the other backup BIOS chip from what new version you must recently updated onto your primary chip. Not sure though. So you will have to look into it yourself. And as you say... its still not supposed to brick both of them.
  2. dreamcat4

    [pre-release] macOS High Sierra

    Hi there. Have any of you guys seen new support added for polaris12 (rx550) in latest DP? Its probably not / no. ​ From 2 months ago: ​ ​> There is no 699F device ID in the kexts to allow Polaris 12 work in Mac OS. ​ ​I just ask because the polaris12 has fewer CUs than rx460. So when we tried to spoof it today (on sierra 12.6) to be rx460, it just hard resets during boot time. Maybe its trying to address processing units which are not there, and then crashes. Anyhow this is not a problem for me. Just thought to ask while still have the card in hands. In few days / couple of weeks I will return this card back to retailer anyhow due to fan noise issue / coil whine. And then revert back to passive gt1030 instead. ​ ​What confused me was I bought this card (not checking carefuly!), just assuming it was already supported. Since a lot earlier on there was a news finding stuff like Polaris12 id in kexts: ​ ​https://forums.appleinsider.com/discussion/197513/trio-of-new-amd-gpu-references-for-possible-mac-refresh-found-in-latest-macos-sierra-beta ​ ​Ah well.
  3. dreamcat4

    [pre-release] macOS High Sierra

    ​Yeah i have an ASUS BT400 too here. And its a very good product, although much more expensive than a BT CSR4 module which is like... only $2 on ebay. So I would recommend just buying that anyhow first of all. Then if it don't work, you've given it a shot but only lost $2 there. Which is literally nothing. Then buy the ASUS BT400.
  4. dreamcat4

    [pre-release] macOS High Sierra

    Are you really so sure about that with the current latest versions of WhateverGreen? Because its FAQ is saying "10.1 and higher". https://github.com/vit9696/WhateverGreen/blob/master/Manual/FAQ.en.md It does not expilitly state which specific features work [or do not work] on specific versions of macOS. Maybe vit9696 the plugin author can clarify that. And has information somewhere else online about it
  5. Yeah... I was about to say: that is clearly a symptom of flash-based devices. If your speed initially starts looking OK, then that means the USB 3 bus speeds is fine. And can handle that bandwidth - all is OK with the USB 3 chipset performance.
  6. Heres my test result on a fat32 partitioned 850 evo ssd, attached on: usb 3.0 enclosure, on usb 3.0 header next to the 24-pin atx socket. ga-z170x-g3:250_ id$ dd if=/dev/zero of=00 bs=1m count=100 100+0 records in 100+0 records out 104857600 bytes transferred in 0.388728 secs (269745328 bytes/sec) ga-z170x-g3:250_ id$ dd if=/dev/zero of=00 bs=1m count=200 200+0 records in 200+0 records out 209715200 bytes transferred in 0.818502 secs (256218327 bytes/sec) ga-z170x-g3:250_ id$ dd if=/dev/zero of=00 bs=1m count=300 300+0 records in 300+0 records out 314572800 bytes transferred in 1.317064 secs (238843966 bytes/sec) ga-z170x-g3:250_ id$ dd if=/dev/zero of=00 bs=1m count=400 400+0 records in 400+0 records out 419430400 bytes transferred in 1.708555 secs (245488383 bytes/sec) and heres the same tests repeated on the same disk as a raw block device (rdisk) ga-z170x-g3:/ root# dd if=/dev/zero of=/dev/rdisk3 bs=1m count=100 100+0 records in 100+0 records out 104857600 bytes transferred in 0.445748 secs (235239720 bytes/sec) ga-z170x-g3:/ root# dd if=/dev/zero of=/dev/rdisk3 bs=1m count=200 200+0 records in 200+0 records out 209715200 bytes transferred in 0.887426 secs (236318484 bytes/sec) ga-z170x-g3:/ root# dd if=/dev/zero of=/dev/rdisk3 bs=1m count=300 300+0 records in 300+0 records out 314572800 bytes transferred in 1.331839 secs (236194342 bytes/sec) ga-z170x-g3:/ root# dd if=/dev/zero of=/dev/rdisk3 bs=1m count=400 400+0 records in 400+0 records out 419430400 bytes transferred in 1.775635 secs (236214312 bytes/sec) ga-z170x-g3:/ root# And in BlackMagic Speed Test.app, it says "225.2 mb/sec write speed, 257.2 mb/sec read speed" Which matches closely with the results from dd command line (225.27 mb/sec for the very last 400mb test here). From memory (i had posted the result in another thread a long time ago). Compared to previous test before, and on this 10.12.6 update now, it seems no different for me cyberdevs. If you are experiencing slower performance now, it might be something else. For example: I had this other SSD here got overloaded too much, it had not enought time to do its garbage collection (to recover itself). And was temporarily in a slow performance mode. Some dude told me to write 000000 the whole drive, and just leave it plugged in idle (powered on system). My SSD came back to its normal performance.
  7. dreamcat4

    [pre-release] macOS High Sierra

    Thank you. ​ ​ ​Question: Is it safe now to do this? I mean because it was a 15 port limit because some potential crash / or overflow in apple XHCI usb driver. Does this fix / avoid that crash overflow (above 15 ports used) ? I ask Rehabman about this a while back. But at that earlier time he said there was no known fix (it being an issue in apple's code).... Yes I know its still ok that many people who just use USBInjectAll.kext and do not bother to worry. Who dont disable any USB ports. And those other guys who have NUC or laptop with much fewer USB ports = less than 15 total. It doesnt matter for those guys.
  8. ^^ This. Aparrently... there is a way to downgrade the BIOS to before F20. But its not so nice and easy thing. Perhaps a little risky to brick the BIOS if you do it wrong. Or else maybe you can use a SOIC8 clip / spi hardware flasher. To bypass the QFlash restriction and manually reflash. Which is another more difficult / harder ways. The other reason to go back is for bclk overclocking of locked skylake CPUs... which disables IGP at OS load time. But you can still on IGP graphics see the boot screen and enter the BIOS screens. I tested / tried this on Gaming 3. ​ ​Again thanks for letting us know! Because I didnt know this either. Using the same method here. But I'm talking to another guy here who has made DSDT for Gaming 3 with other USB3 method (not USBInjectAll.kext)... so I will try test / compare tomorrow also. (both kinds his way and mine). ​ remember: ​For testing USB 3 speeds, its... dd if=/dev/rdiskX of=/dev/null bs=1m count=100 the device name in OS X must be the /dev/rdiskX one and not /dev/diskX OK guys! ​
  9. Hello again cyberdevs, I was about to try this Kext, however then I tested before to find the iBooks transparent window has already been fixed now. That is in the newest 10.12.6, with the newest .25. nvidia web driver. So it seems this kext is not needed [for that] anymore. However readme on github does say 'makes nvidia the primary graphics'... So perhaps there can still be other graphics issues which this kext might be able to solve. Dont know... Anyhow the only glitch i saw anymore is: black rectangle in the system tray (menubar). So if it solves other glitches then its still useful i guess? So I'm going to hang on to this one. But not going to apply yet until I first see something else to reproduce and test it with / without the kext.
  10. ​Thanks for letting us know... I suppose we can just put in our clover kexts folder like the others.
  11. err.. maybe try turning off 'power nap'. see if that helps.
  12. CUDA question? Hello! Did not know until reading your posts ^ above today, that there were existed such CUDA drivers! There is some good benefit to it then? ... is it for rendering tasks in Premiere / Photoshop / other professional apps? Or does this help something else in os x? ==== Gaming 3 updated: Anyhow I have fixed up all my Gaming 3 Clover configs today. Its available upon request (just PM me here). Updated for 12.6 today. I didnt do all this loads of DSDT patching to the max (like you guys for the Gaming 5). So mine has always been broken sleep function, it doesnt matter to me. Everything else seems fine. Many thanks for all help / knowledge coming back here. Just to update the CLOVER again for latest. BTW whatever happened to Gaming 4 guys? Did we skip one? Hehehe. ==== Testing Metal: And something for future: I find a list of perhaps 'metal' games? https://forums.macrumors.com/threads/list-of-mac-metal-games.2045606/ We can try that list to to see which ones are the best for benchmarking metal (if they can at all). So to proper 'decent' testing for latest nvidia web drivers (will be hoping). Actually I already posted on other forum thread about metal games list --> http://www.insanelymac.com/forum/topic/312525-nvidia-web-driver-updates-for-macos-sierra-update-07212017/?p=2467028 But maybe it should get its own thread (not to be lost in massive thread).
  13. Just upgraded to 10.12.6, and the above ^ Nvidia WebDriver-378.05.05.25f01... was not the *smoothest* software update I've ever experienced... but it all seems ok and fine enough in the end. BTW Also reports are that 10.12.6 now accepts iMac 18,[1,2,3] hardware profile in SMBIOS wizard of Clover Configuratior... (though that is nothing to do with these NVIDIA drivers). Anyhow I see above here some talk about retesting webkit? Sorry I don't know how to test for webkit. But I can say it seems now the @ Apple logo is correctly restored in the corner of Safari. Not anymore some strange or corrupted unicode glytph [] symbol ... It will be nice to see if any other graphics glitches dont come back. Actually I can still see remains the black rectangle in menubar glitch. Which is still there. Not gone away yet. But its not appear too much, so is tolerable / not too bad. Anyhow I was trying also to bench nvidia GPU very recently... but unfortunately all the benchmark tools I found were based on OpenGL and OpenCL entirely. BUT! Then I come across this list here: https://forums.macrumors.com/threads/list-of-mac-metal-games.2045606/ Which is supposed to be all metal games. Looks good to me! But unfortunately dont have any time yet to properly check any of them out myself. So it would be great (if any you guys want to try them out / already familiar with some games on this list?)... THen we can find out properly which of these games are the best ones for benchmarking metal. And to re-test the situation for future NVIDIA webdriver updates. Maybe we can refine this list with our own notes and experiences. Or simply just re-order this list from best --> worst. [Purely in terms of just benchmarking for metal API performance, and find the ones that a most *easy* / fast to benchmark.] Looking forward to it.
  14. dreamcat4

    macOS Sierra 10.12.6 final is out.

    OK here are all my most recent benches taken on 10.12.5, with all newest drivers. The unigine results are raw, and not summarized (tableized) in the test-results .txt files. Sorry I didnt have time to properly organise these results. So you have to trawl through some separate files to make a comparison. This applies only to the unigine test results (each test is individual html file). Which are not labelled. https://www.dropbox.com/sh/3in9ny8m2u7jh6e/AADF4k_U6-OwFoW4XVfqk6yKa?dl=0 Notes: 3 different gpus were tested: * nvidia gt 1030 * amd rx460 (NOT unlocked cus) * nvidia gtx 1050 * Some unigine tests (html file) say its 'intel hd530'... actually those are the RX460 results. I didnt test any intel IGP whatsoever. * Also double-benched most of these tests in windows 10 too, for comparison. * Luxmark is v3.0 (not v3.1, because they changed luxmark and are not directly comparable results). * Used iMac 18,2 smbios profile, because that iMac hardware, has a version of the amd rx460 (baffin). If that actually matters i dont know. * You will need to tell me if there is an nvidia web drivers OK for 10.12.6 available yet. This is all for 10.12.5 (very recent, 2 days ago). As dont know anything about this 12.6 update yet. Also if no new nvidia webdrivers then these are really new, and perhaps its not matter re-benchmarking? (if its being the same exact nvidia driver for 12.5 and 12.6?). Please correct me here. * For each test run, only that GPU was physically installed in the pci-e slot. * Did not have access to any metal tests. Its all openGL testing, (+ some microsoft DX on windows). Since cannot login yet, to download gfxbench metal from app store. * Didnt have any time left to read these test results. So comments / analysis are very welcome.
  15. dreamcat4

    macOS Sierra 10.12.6 final is out.

    Hello. Have some recent benchmarks here on 10.12.5, and a pretty new nvidia web drivers version. If you are interested then I will post them later today.
×