You are not logged in.
I have two laptops connected to the same monitor (Dell G3223D) via HDMI. Both laptops run Arch Linux. One laptop has an NVIDIA card (RTX 3080) and the other is an Iris Xe (i7-1260p). The monitor supports 165Hz refresh at 2560x1440, however only 144Hz is available via HDMI, which is fine.
The problem is that the laptop with the NVIDIA card can use 1440p at 120Hz and 144Hz with no issues, but the Iris Xe one can only use 60Hz at 1440p. If I boot the Xe laptop into Windows I can use 144Hz with no issues. Xrandr only reports 59.95Hz for 1440p on the Xe laptop, but on the NVIDIA laptop it reports all 3 refresh rates.
Any ideas?
Offline
Please post the xorg log ( https://wiki.archlinux.org/title/Xorg#General - on the Xe laptop) and see https://wiki.archlinux.org/title/Xrandr … esolutions
Offline
Yes, the Xorg log doesn't show high refresh modelines. I'll try parsing the EDID to see if the resolutions are there, I assume they are not as there are no EDID errors reported in the log. How come the other laptop with the NVIDIA card can detect all modes just fine? Doesn't it also use the same EDID?
By the way, do you know hot to get (some equivalent of) the Xorg log when running GDM and Gnome in Wayland mode? I had to switch to Gnome on Xorg to get the log.
Last edited by borisbsc (2022-09-02 14:13:59)
Offline
I'll try parsing the EDID … I assume … How come
Please post the xorg log
You're asking to answer questions on the base of no data.
How come the other laptop with the NVIDIA card can detect all modes just fine? Doesn't it also use the same EDID?
Translates to
Doesn't [the other laptop] also use the same EDID?
Answer implied.
By the way, do you know hot to get (some equivalent of) the Xorg log when running GDM and Gnome in Wayland mode?
Such doesn't exist - but gdm/gnome logs everything into the system journal (sudo journalctl -b)
I had to switch to Gnome on Xorg to get the log.
Xrandr only reports 59.95Hz for 1440p on the Xe laptop
xrandr outputs under wayland are use- and meaningless.
xrandr is at this point talking to a nested X11 server, kind of a virtual machine (not really, but I hope you get the point)
On a random guess this could be https://wiki.archlinux.org/title/Variable_refresh_rate related
Offline
I updated the original post with an Xorg log upload but you were too quick to answer ![]()
Here it is again: Xorg log
Btw I was using Xfce when I originally ran xrandr and posted the question, I have switched to Gnome since. Thank you for the answer regarding Wayland logs, didn't know that. I'll try comparing the EDID/Xorg log from the NVIDIA laptop and I'll try forcing modes with xrandr in xorg mode to see if it works. If it does I still have no idea how to force Wayland into specific display timing, maybe I can force the EDID somehow via kernel module or smth.
Offline
xrandr --verbose on NVIDIA
xrandr --verbose on Iris Xe
EDIDs are obviously different. How come, a bug in the intel driver or using different methods for obtaining the EDID?
Last edited by borisbsc (2022-09-02 14:46:09)
Offline
There're two EDIDs, one is the internal display (sorry, I forgot that this was about an external one - yeah, should end up w/ the same EDID file) by SDC
DTD 1: 2880x1800 90.001438 Hz 8:5 169.203 kHz 500.840000 MHz (302 mm x 189 mm)
Hfront 8 Hsync 32 Hback 40 Hpol P
Vfront 8 Vsync 8 Vback 64 Vpol N
DTD 2: 2880x1800 60.000958 Hz 8:5 169.203 kHz 500.840000 MHz (302 mm x 189 mm)
Hfront 8 Hsync 32 Hback 40 Hpol P
Vfront 8 Vsync 8 Vback 1004 Vpol NThe other one is a DELL
Block 0, Base EDID:
DTD 1: 2560x1440 59.950550 Hz 16:9 88.787 kHz 241.500000 MHz (697 mm x 392 mm)
Hfront 48 Hsync 32 Hback 80 Hpol P
Vfront 3 Vsync 5 Vback 33 Vpol N
Block 1, CTA-861 Extension Block:
DTD 2: 2560x1440 143.911545 Hz 16:9 222.056 kHz 592.000000 MHz (697 mm x 392 mm)
Hfront 8 Hsync 32 Hback 66 Hpol P
Vfront 25 Vsync 8 Vback 70 Vpol N
DTD 3: 2560x1440 119.997589 Hz 16:9 182.996 kHz 497.750000 MHz (797 mm x 334 mm)
Hfront 48 Hsync 32 Hback 80 Hpol P
Vfront 3 Vsync 5 Vback 77 Vpol NBut
modeset(0): HDMI max TMDS frequency 300000KHz looks like this is HDMI 1.3/1.4 only, which is not enough (and the supported modes stay right below that threshold)
Do you have specs for the notebook that suggest otherwise?
Offline
Yeah, the eDP is the built-in display and HDMI-1 is the external monitor. HDMI version is 2.0, besides, it all works in Windows. I updated the previous post with
xrandr --verbosefrom both laptops, and EDIDs are obviously different.
xrandr --verbose on NVIDIA
xrandr --verbose on Iris Xe
Last edited by borisbsc (2022-09-02 14:49:38)
Offline
@@ -24,7 +24,7 @@ Block 0, Base EDID:
EDID Structure Version & Revision: 1.3
Vendor & Product Identification:
Manufacturer: DEL
- Model: 53557
+ Model: 53553
Serial Number: 810234188
Made in: week 1 of 2022
Basic Display Parameters & Features:
@@ -66,7 +66,7 @@ Block 0, Base EDID:
Display Range Limits:
Monitor ranges (GTF): 48-144 Hz V, 30-230 kHz H, max dotclock 600 MHz
Extension blocks: 1
-Checksum: 0xd9
+Checksum: 0xdd
----------------
@@ -105,7 +105,7 @@ Block 1, CTA-861 Extension Block:
IT scan behavior: Always Underscanned
CE scan behavior: Always Underscanned
Vendor-Specific Data Block (HDMI), OUI 00-0C-03:
- Source physical address: 2.0.0.0
+ Source physical address: 1.0.0.0
DC_36bit
DC_30bit
DC_Y444
@@ -144,4 +144,4 @@ Block 1, CTA-861 Extension Block:
DTD 3: 2560x1440 119.997589 Hz 16:9 182.996 kHz 497.750000 MHz (797 mm x 334 mm)
Hfront 48 Hsync 32 Hback 80 Hpol P
Vfront 3 Vsync 5 Vback 77 Vpol Ndecoded diff - no relevant difference.
https://wiki.archlinux.org/title/Xrandr … esolutions
% cvt12 2560 1440 120 -b
# 2560x1440 @ 120.000 Hz Reduced Blank (CVT) field rate 120.000 Hz; hsync: 183.000 kHz; pclk: 483.12 MHz
Modeline "2560x1440_120.00_rb2" 483.12 2560 2568 2600 2640 1440 1511 1519 1525 +hsync -vsync
% cvt12 2560 1440 144 -b
# 2560x1440 @ 144.000 Hz Reduced Blank (CVT) field rate 144.000 Hz; hsync: 222.192 kHz; pclk: 586.59 MHz
Modeline "2560x1440_144.00_rb2" 586.59 2560 2568 2600 2640 1440 1529 1537 1543 +hsync -vsyncThe modelines w/ full blanking exceed 600MHz (and are not supported by HDMI 2.0, but would be by HDMI 2.1
Edit: nvidia btw runs w/
cvt12 2560 1440 120 -r
# 2560x1440 @ 120.000 Hz Reduced Blank (CVT) field rate 119.998 Hz; hsync: 182.996 kHz; pclk: 497.75 MHz
Modeline "2560x1440_120.00_rb1" 497.75 2560 2608 2640 2720 1440 1443 1448 1525 +hsync -vsyncAnd (this is not the cvt12 reduced blanking, but something from the edid)
# 2560x1440 @ 144.000 Hz (from the edid and nvidia) field rate 143.91 Hz; hsync: 222.06 kHz; pclk: 586.59 MHz
Modeline "2560x1440_144.00_nvidia" 592.00 2560 2568 2600 2666 1440 1465 1473 1543 +hsync -vsyncLast edited by seth (2022-09-02 15:21:28)
Offline
I added a new mode based on NVIDIA-reported timings:
$ xrandr --newmode "2560x1440_120.00" 497.75 2560 2608 2640 2720 1440 1443 1448 1525 +hsync -vsync
$ xrandr --addmode HDMI-1 2560x1440_120.00Verified that it's added:
$ xrandr --verbose
2560x1440_120.00 (0x82f) 497.750MHz +HSync -VSync
h: width 2560 start 2608 end 2640 total 2720 skew 0 clock 183.00KHz
v: height 1440 start 1443 end 1448 total 1525 clock 120.00HzIt exactly matches the NVIDIA timings.
However, running `xrandr --output HDMI-1 --mode 2560x1440_120.00` produces:
xrandr: Configure crtc 0 failedI'm out of ideas. Maybe the HDMI port is being detected as < 2.0, which would limit the pixel clock to 340 MHz (1440p@120 is 497.75 MHz). If it didn't work in Windows I would assume the cable is at fault or that the HDMI port on a 2022 laptop is really 1.4b.
EDIT
Yes, you noticed previously:
modeset(0): HDMI max TMDS frequency 300000KHz^ This seems to be the problem. Why is the TMDS frequency so low? Both the port and cable are HDMI 2.0, can I force it somehow?
Last edited by borisbsc (2022-09-02 15:50:26)
Offline
You could try the xf86-video-intel driver.
Can you test the behavior w/ a displayport connection?
Offline
Tried xf86-video-intel, other than some graphical glitches and jerkiness the max refresh rate reported is still 60Hz.
USB-C (DP-1) works, at 165Hz too. I guess I need to wait for a kernel update or something.
Offtopic - I'm avoiding USB-C as there seems to be some hardware defect with this particular unit. When connected via USB-C the two USB-A ports on the monitor front are powered on, however using them either doesn't work at all or Windows keep reporting USB power surge warning every 5 seconds (and with a trivial device plugged in, like a keyboard/mouse dongle). Also, the 3.5mm headphone jack that sits right between the two USB-A ports continuously emits strong, loud pulses that almost made my ears bleed and nearly ripped the speaker membrane on my headphones. On the other hand, the unit has zero bright or dark pixels and absolutely no backlight bleed and glow, so I'm too anxious to return it under warranty and replay the panel lottery... ![]()
Offline
https://wiki.archlinux.org/title/intel_ … nable_fbc) (but I'd not hold my breath for this to help)
If you can get hands on other (higher grade) HDMI cables, try them - HDMI sucks™, the handshake is hyper-fragile and the cables have an active part in this ![]()
Offline
Yup, tried that already. Fastboot and PSR too, no joy. Thank you anyway for walking through this with me today, I learned a lot.
Offline