Nvidia Geforce 416.16 WHQL - Download & Discussion

Lol bs. Its on my current z87 with native pcie3.

It's a asus deluxe variant for 250€ that is nothing cheap about it. And you don't know anything. Just acting smart that's all.




Anyway why is this offtopic still going? You are all Better off with other more sensible optimizaions.

//Just saying.

 

you are wrong for example in x58 motherboards when the driver use msi mode it use interrupts directly on cpu when the driver use legacy irq it reroutes the fuction to ich10r

 

*nvm

 

I am right
https://www.oreilly.com/library/view/pci-express-system/0321156307/0321156307_ch09lev1sec5.html

MSI's are a DMA technique, for which the root ports are configured to msi's even when the card in the slot is not, you are never "rerouted to the ich10r"

If your Root ports are configured to IRQ's, something is wrong with your system install because these are configured to MSI's by the driver built into windows.

x58's MSI capability is described in detail
https://www.intel.com.au/content/dam/doc/datasheet/x58-express-chipset-datasheet.pdf

HwInfo can be wrong you know.

The asus deluxe at the time was a cheap designed poorly implemented board sold at a premium, it had several design issues that were addressed in the P6X58 and RoG boards

It's also not off topic, this is discussion of the drivers.

 

you are wrong. Interrupts handled from ich10r chip set. when irq have a positive value.

 

No, sorry but you are wrong and have misread the chipset datasheet.

Interrupts generated by a source on the IOH processed then passed untouched to the CPU via QPI link as an MSI.

Interrupts fired from a PCI or PCI-X device on the ICH are however, handled by fixed range decoders in the IOH.

See sections 8.2.3 and 8.3

It actually seems you did see these, and have mis-interpretted the section detailing how a GPE event is handled when a hotplug event occurs on a PCX device device.

But 8.4 and onwards completely invalidates your claims on IRQ rerouting.

infact, 8.1, and 8.1.2 completely backs up what i have said, and you are claiming is wrong

There is only one case in which Interrupts are forwarded to the ICH, and This circumstance is not possible on a multicore cpu as it involves returning to the single processor PIC kernel,

In all cases of a multicore functioning x58 setup, the IO Apic is enabled, Interrupts are converted to native MSI's.

 

Commenting someone's hardware is "cheap" is off topic...

 

fact is, it wasn't MSI that caused the stutters, if anythig it was the nv gen3 tool being unstable, assuming it was the x79 deluxe and not the x58 deluxe board from asus at this point.

MSI's are always used on PCI-E APIC controllers.

 

x58 have 2 steppings. b2 (12) and b3 ( 13) in b2 it have ioapic disabled in b3 it have ioapic enabled.
-https://lists.openwall.net/linux-kernel/2009/09/07/8

 

you cannot use MSI's at all without IOAPIC.

If your board goes into MSI, you don't have a B2.

AIDA and HWInfo will tell you what stepping your X58 is.

There is also a C2 stepping floating about.

 

I have the B3 revision of X58 on my board. Some days ago I made the switch from the GTX 670 to an RX 570, the AMD driver by default uses MSI mode and I get also a BSOD (WHEA uncorrectable, Bugcheck 124) in Battlefield 1 after 1+ hours of gameplay. Setting the card up in IRQ mode also gives me a smoother image and will test today if the Bluescreens are gone or not. So this could be a Microsoft issue after all - I didn't have these problems before 1809 and usually MSI mode gave me better results in smoothness.

 

im running 1809 on my R2G, with MSI's and there are no bsods with latest drivers (which, i should note is a B2/Rev12)

It could just be a radeon driver bug, or one of the issues that 17763.104 was just released to address in the insider program.

 

Well, it seemed a bit odd that both cards from different vendors show the same symptoms which lead me to believe that there could be a different source for this error. The error code usually points towards a CPU voltage problem but my settings were fine earlier and could stand a Linux Kernel compilation from source which should stress the system enough to rule out that possibility. Also seeing other users over here complaining about this BSOD, makes me confident that this is not a hardware failure on my system. I tested BF1 this afternoon and the system was stable. Hence setting up the card in IRQ mode might be the way to go on my X58 system with the current build of 1809. I'll certainly try the new update when it becomes available, maybe even the new Insider Preview.

 

Just keep it at default, problem solved. Well if you set audio card e.g. xfi pcie to Msi it stops working.

How I have it set, use to be more components by older windows, now it's almost half less.
https://i.**********/cKBR0pyH/msi.png


Anyway, those who think this will fix stutters, sorry no. Your problem lies elsewhere.

 

thats because the xfi hardware design operates in 2 modes, and the full - creative driver mode operates multiple kernel drivers which only communicate through interrupts on the device.

Everything consisting of an interrupt beyond the cards slot is an MSI via the IOAPIC.

nvidia's ISR's are reduced with MSI's, but DPC latencies are still at the mercy of accumulated results, with the intel nic in your case more often then not spiking the rest of the system.