Contribute
Register

Report your NVIDIA Web Driver and CUDA issues and I'll pass them on to the developers

Status
Not open for further replies.
I'm not telling you that Nvidia won't come up with some update about this issue we're facing, but it will be pretty unclear if they actually will care about it, since they're looking forward the new releases.

I tried everything in this issue. I'll post my configuration here, 'cause I'm didn't update my profile yet, but my specs are

- mobo: Asus B150M C/BR
- CPU: i5 7400 3.0Ghz - 3.5Ghz
- RAM: Hyper X 8GB 2333Mhz
- GPU: Nvidia 1050TI 4gb
- 2 monitors: LG E2350 and MP55

- High Sierra 10.3.3

Well, I tried Lilu.kext, graphicsFixup, whateverGreen ( seemed to correct the issue introduced by Metal2 ), changed CPUFakeId, Shiki, FrameBuffer, change NVCap, set VRAM manually, correct the ig-platform-id, change SMBIOS, and many, many other fixes

I was really confident that it was my fault into new installation that caused the lagginess. And, I'll confess, this lags made me so angry that for 1 month I switched to Windows almost completely.

The only solution for me, at least, was downgrade my Nvidia Web Driver. There was so much I was been through that I didn't care at the time that would broke up my whole install, but surprisingly that worked PRETTY WELL.

All my GUI animations were working flawlessly again and I felt again like I was in MacOS. I have noticed that Videos in iTunes and RAM consumption where normalised when I was watching a movie and youtube wasn't laggy anymore

I really don't know if this link that I am about to drop here violates the tonymac's policies, if it do so, please, tell me and I'll remove immediately ( I really do not want my account banned, I just want to help )

In this link you'll find a guide that can help you with your issue, installing a downgraded driver for your Nvidia web driver
https://www.******.com/r/hackintosh/comments/7sr4vv/nvidia_web_drivers_and_you_a_patching_guide_for/
 
A user reported on my Z370-G build thread that he is not experiencing "the lag" here. What I found interesting was that he disabled two PCI-e slots in BIOS. Someone (can't remember who) previously had mentioned that the lag seemed more severe with lots of network activity.

Perhaps the issue is related PCI lanes. Maybe we can eliminate the lag by disabling some devices in BIOS...
 
A user reported on my Z370-G build thread that he is not experiencing "the lag" here. What I found interesting was that he disabled two PCI-e slots in BIOS. Someone (can't remember who) previously had mentioned that the lag seemed more severe with lots of network activity.

Perhaps the issue is related PCI lanes. Maybe we can eliminate the lag by disabling some devices in BIOS...

I agree that PCIe is taxed on 100/200/300-series. Mainstream boards really should have more lanes at this point. I couldn't install my TB3 add-in card even tho according to the manual I should have enough lanes. I had to choose between it and the SATA controller and I chose SATA controller. Yes, disabling individual SATA ports didn't work, only the whole controller. And the same issue happened in Windows. The PCH would fault and kill all my USB ports until I power-cycle. The manual did not mention this issue. :/

At the same time, I don't think this is the issue with the 156/157 driver. The lag is just too widespread. Maybe the 156/157 drivers are capping the bandwidth at 4x internally somehow, but I doubt that disabling slots/lanes/ports will really help. Issues with network activity are consistent with what I observed with IO and suggest maybe it's related to some new Spectre/Meltdown mitigations in the driver . . .
 
Sorry if I missed it, but how does this ”lag” you speak of manifest itself? Is it noticeable when moving a window around or what is the best way to check if I have the ”lag” issue?
 
Last edited:
Sorry if I missed it, but how does this ”lag” you speak of manifest itself? Is it noticable when moving a window around or that is the best way to check if I have the ”lag” issue?

If you have "the lag", there's nothing you have to do. It will be immediately apparent the second you try to do anything in Finder. Simple things like maximizing a window to full screen will lag.

Odds are you will not see this since you are on an i7-920. As I've stated many times, this issue seems to primarily only affect Skylake and newer systems.
 
I agree that PCIe is taxed on 100/200/300-series. Mainstream boards really should have more lanes at this point. I couldn't install my TB3 add-in card even tho according to the manual I should have enough lanes. I had to choose between it and the SATA controller and I chose SATA controller. Yes, disabling individual SATA ports didn't work, only the whole controller. And the same issue happened in Windows. The PCH would fault and kill all my USB ports until I power-cycle. The manual did not mention this issue. :/

At the same time, I don't think this is the issue with the 156/157 driver. The lag is just too widespread. Maybe the 156/157 drivers are capping the bandwidth at 4x internally somehow, but I doubt that disabling slots/lanes/ports will really help. Issues with network activity are consistent with what I observed with IO and suggest maybe it's related to some new Spectre/Meltdown mitigations in the driver . . .

I tested my theory about the lag being PCI lane related.
  • I took out my 10GBase-T Card
  • Disabled USB 3.1 Gen 2
  • Disabled Wi-Fi
  • Disabled Bluetooth
  • Disabled Built-in LAN
My graphics card was running at x16. No difference. Lag still persists... :banghead:
 
I tested my theory about the lag being PCI lane related.
  • I took out my 10GBase-T Card
  • Disabled USB 3.1 Gen 2
  • Disabled Wi-Fi
  • Disabled Bluetooth
  • Disabled Built-in LAN
My graphics card was running at x16. No difference. Lag still persists... :banghead:


I just tested on my new system with 44 Lanes x299 CPU....same lag.

Put back in the RX560 and will wait until nVidia and Apple fixes this.

If you go back to Sierra 10.12.6 you will see even on a 6700k which doesn't have as many PCIe lanes as an HEDT CPU, it runs like butter even on an older Maxwell 980Ti.
 
Potential cause and possible fix for the lag.
As Firefox doesn't display the same lag as Safari, at least on YouTube, I believe the lag might be caused by Shiki not using the proper patches for Kaby Lake. I think this part from Patches.plist is not properly patched for iMac18,2
Code:
                    <key>Comment</key>
                    <string>board-id -&gt; shiki-id (normally for iMac14,2)</string>
                    <key>CPU</key>
                    <string>CPU_TYPE_X86_64</string>
                    <key>Find</key>
                    <data>Ym9hcmQtaWQA</data>
                    <key>Replace</key>
                    <data>c2hpa2ktaWQA</data>
                    <key>Skip</key>
                    <integer>0</integer>
                    <key>Count</key>
                    <integer>1</integer>
                    <key>Section</key>
                    <string>BOARDID</string>
                    <key>Segment</key>
                    <string>TextCstring</string>

And the difference between Firefox
Code:
            <key>Comment</key>
            <string>firefox and firefox-bin</string>
            <key>Path</key>
            <string>/Firefox.app/Contents/MacOS/firefox</string>
            <key>Section</key>
            <string>NVDA</string>
            <key>Flags</key>
            <string>PF::MatchAny</string>
and Safari
Code:
            <key>Path</key>
            <string>/Applications/Safari.app/Contents/MacOS/Safari</string>
            <key>Section</key>
            <string>NVDA</string
 
Last edited by a moderator:
Status
Not open for further replies.
Back
Top