> > In the 2.1.117 kernel, we have the mtrr (memory type range register)
> > support. When we activated the mtrr for the video card at the prompt
> > (by following the instructions of the mtrr documentation), things got
> > even worse, performances dropped by about 50% ! According to the mtrr
> > documentation we should have seen improvements. We also tried mtrr
> > with a dual IBM Intellistation Z Pro equipped with 2 Pentium Pro 200
> > MHz 512 kb cache running kernel 2.1.115 and, again, we got a
> > performance degradation when we enabled mtrr.
>
> Bizarre. I can't think how the MTRR stuff could affect the
> non-graphical performance to any significant extent (in either
> direction), unless it was configured in ways that shouldn't be
> possible with the current MTRR driver.
>
> Do you get a performance degradation between a kernel built without
> CONFIG_MTRR and one with, or does it only appear after you write to
> /proc/mtrr to configure it for your video card?
>
> Could you post the output of "cat /proc/mtrr", and the output of
> "scanpci" or "lspci", please.
Hi,
we tried the same test with a dual PPro 200 MHz (PCI Matrox
Millenium I with 4 Mb RAM) running kernel 2.1.119. This time,
when we enabled the mtrr we did not see any significant
performance modifications. The timings were unchanged with or
without mtrr enabled.
As requested, here is the output of "cat /proc/mtrr":
======================================================
reg00: base=0x00000000 ( 0MB), size= 128MB: write-back, count=1
reg01: base=0xc0000000 (3072MB), size= 4MB: write-combining, count=1
And the output of "scanpci":
============================
PCI says configuration type 1
PCI probing configuration type 1
Probing for devices on PCI bus 0:
pci bus 0x0 cardnum 0x00 function 0x0000: vendor 0x8086 device 0x1237
Intel Device unknown
pci bus 0x0 cardnum 0x01 function 0x0000: vendor 0x8086 device 0x7000
Intel 82371 pci-isa bridge
pci bus 0x0 cardnum 0x01 function 0x0001: vendor 0x8086 device 0x7010
Intel 82371 bus-master IDE controller
pci bus 0x0 cardnum 0x01 function 0x0002: vendor 0x8086 device 0x7020
Intel Device unknown
pci bus 0x0 cardnum 0x02 function 0x0000: vendor 0x1011 device 0x0024
Digital Equipment Corporation DC21152
pci bus 0x0 cardnum 0x0f function 0x0000: vendor 0x9004 device 0x8178
Adaptec 2940U
pci bus 0x0 cardnum 0x10 function 0x0000: vendor 0x102b device 0x0519
Matrox MGA Millennium
Probing for devices on PCI bus 1:
pci bus 0x1 cardnum 0x02 function 0x0000: vendor 0x8086 device 0x1229
Intel 82557 10/100MBit network controller
and finally, a small part of the "startx" output"
=================================================
<cut>
(--) SVGA: PCI: Matrox MGA 2064W rev 1, Memory @ 0x80000000, 0xc0000000
(--) SVGA: Linear framebuffer at 0xC0000000
(--) SVGA: MMIO registers at 0x80000000
(--) SVGA: Video BIOS info block at 0x000c7bfc
(--) SVGA: chipset: mga2064w
(--) SVGA: videoram: 4096k
(**) SVGA: Option "dac_8_bit"
(**) SVGA: Using 8 bits per color component
(**) SVGA: Using 8 bpp, Depth 8, Color weight: 888
(--) SVGA: Maximum allowed dot-clock: 220.000 MHz
(**) SVGA: Mode "1024x768": mode clock = 85.000
(**) SVGA: Virtual resolution set to 1024x768
(--) SVGA: Generic SpeedUps selected (Flags=0x30)
(--) SVGA: MCLK set to 50.000 MHz
(--) SVGA: Using hardware cursor
(--) SVGA: Using XAA (XFree86 Acceleration Architecture)
(--) SVGA: XAA: Solid filled rectangles
(--) SVGA: XAA: Screen-to-screen copy
(--) SVGA: XAA: 8x8 color expand pattern fill
(--) SVGA: XAA: CPU to screen color expansion (TE/NonTE imagetext, TE/NonTE polytext)
(--) SVGA: XAA: Using 8 128x128 areas for pixmap caching
(--) SVGA: XAA: Caching tiles and stipples
(--) SVGA: XAA: General lines and segments
(--) SVGA: XAA: Dashed lines and segments
<cut>
==================================================
Thank you for your help, we greatly appreciate,
Claude
-- Claude Gamache, CAE Electronique Ltee, 8585 Cote-de-Liesse Saint-Laurent, Quebec, Canada H4T 1G6 Email: cgamache@cae.ca Tel.: (514) 341-2000 x3194- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.altern.org/andrebalsa/doc/lkml-faq.html