Re: AMD X2 unsynced TSC fix?

On Sat, Oct 28, 2006 at 12:18:00PM -0700, thockin@xxxxxxxxxx wrote:
On Sat, Oct 28, 2006 at 09:15:15PM +0200, Willy Tarreau wrote:
While gtod is time critical and often appears high on profile lists it is
normally not as time critical as you're claiming it is; especially not
time critical enough to warrant such radical action.

Yes it was, because the small gain of using a dual core with such
a workload was clearly lost by that change. IIRC, I reached 25000
sessions/s on dual core with TSC if I didn't care about the clock,
20000 without TSC, and 18000 on single core+TSC. But with the sniffer,
it was even worse : I had 500 kpps in dual-core+TSC, 70kpps without
TSC and 300 kpps with single-core+TSC. Since I had to buy the same
machines for both uses, this last argument was enough for me to stick
to a single core.

Was the problem that they were not synced at poweron or that they would
drift due to power-states?

They resynced at power up, but would constantly drift. I don't even know
if it was caused by power states. When the machine was loaded, a single
task moving across the cores could see its time jump back and forth
several times a second by an offset sometimes close to +2/-2s.

Did you try running with idle=poll, to avoid ever entering C1 state (hlt)?

Yes, I remember trying such things. I also tried 'nohlt', completely
disabling power management, including ACPI, etc... I also tried vanilla
kernels as well as severely patched ones, but the problem remained the
same in all circumstances, that only 'notsc' could solve.

BTW, I've just found a remain of dmesg capture after boot in case you'd
like to look for anything in it.

Regards,
Willy

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Relevant Pages

  • Re: Intel "speedup" IS mostly RDF
    ... >> I couldn't find anything about the Mobility Radeon x1600's power ... > low-power leader, the Pentium M, because Intel publishes thermal design ... specifications for the dual-core G5 chips used in the Power ... > intel iMac has an aggregate score of 108; the dual core g5 desktop, ...
    (comp.sys.mac.system)
  • Re: Athlon64 X2 3800+ Feedback need I WANT one but is it good ??
    ... >> when compared to the current version of the single-core ... you aren't comparing chips with the same ... And an Athlon 3800 is not exactly a low power chip anyways. ... >don't see dual core processors at all being useful for over a year or more, ...
    (comp.sys.ibm.pc.games.action)
  • Re: AMD X2 unsynced TSC fix?
    ... sessions/s on dual core with TSC if I didn't care about the clock, ... Again most workloads are not that gtod intensive. ... that a per-CPU TSC could help a lot at getting monotonic clocks. ...
    (Linux-Kernel)
  • Re: problem with rdtsc on Pentium M
    ... The specifications allowed this since the introduction of the tsc. ... It counts actual cpu cycles and not time. ... settings automatically depending on the power state of the computer. ... My advice would be to always use external reference timers. ...
    (comp.lang.asm.x86)
  • Re: AMD Quad Core clock problem?
    ... Please also describe the drift quantitatively. ... the early dual core days when this took months to get ... With the older chips, each core had its own TSC, which caused synchronization problems. ... The Barcelona generation chips have a constant frequency TSC on the northbridge, so they should be immune to these problems. ...
    (Linux-Kernel)