Hi again,
Status: Almost perfect now, occasional "code 19" halts (rarely) - not due to the workload, something else weird is going on.
It seems to me that after a "code 19" halt, even though the system has been reset, the R4E is picking up previous "code 19" trail on startup and doesn't really start fresh.
This leads to repeating boot (POST) code 19 problems, even though when the system once has started it seems to run rock solid now (e.g. full load LinX without any sign of issues etc.).
Actions taken thus far:
WAVE 1
1. Started with a "clean" table: Reloaded default BIOS (F5) -> save & reset (F10). Note that the CPU is overclocked by default! (target 3900G)
2. Disabled all unused components/features (e.g. BT, Asmedia OPROM, etc.) -> save & reset
3. Set OC mode to manual and typed in the Kingston XMP values manually (VCCSA left to Auto!) -> save & reset
Gave it one pass with Memtest86+ (150 minutes) = 0 errors (Memtest86+ version 4.20)
On exit from Memtest86+ got "code 19" halt
Restarted the rig and it stopped on BIOS setup with a message: "Overclocking failed..."
Observed that the BIOS was pushing DDR3-1333 mode, rather than the manually set target DDR3-1600 (XMP)
WAVE 2
1. Increased the VCCSA voltage stepwise (0.05v increments) to 1.05v -> save & reset several times = all Ok
2. Started Windows and ran various stress tests (AIDA64, SiSandra, LinX), repeatedly - everything Ok
3. At some point Windows crashed on starting AIDA64 (when loading BIOS data) -> rebooting halted on "code 19" again
This time R4E went into a real "code 19" cycle, and it was very hard to get out of it. It seems that crashing Windows is causing harder R4E hangups.
Eventually the boot moved on and halted on code 94, and then finally next reboot went back to a working system.
I was curious so without changing anything else (other than the VCCSA=1.05v adjustment), I gave this setup another pass with Memtest86+ = 0 errors!
Basically, Memtest86+ has never complained about the memory or the way it works. Something else is going on...
Indeed, on exiting Memtest86+ I got the "code 19" halt again.
However, rebooting the system went fine this time, and I guess R4E just don't like the way Memtest86+ is terminating
Note that I've been running all the way with memory in a DDR3-1600 mode (checked regularly with CPU-Z)
WAVE 3
I decided to test the rig with a heavy workload, just to see if stressing the system is what is causing these problems, or there's something else.
1. Started with running AIDA64 stability stresstest for 30 minutes - no problems and everythings was Ok during
2. Ran full load LinX 0.6.4 for 30 minutes - no problems whatsoever (I was actually able to run other programs on top, the system didn't get sluggish at all!)
3. I went thru all the SiSandra queries collecting various HW information - worked like a charm
So at this point, I'm convinced that the RAM in itself is Ok, once started it's working fine, while "code 19" halts appear (now only occasionally) in various rebooting scenarios.
Finally, I entered 1.15v for VCCSA and 10 for CAS and tRCD as recommended, but unfortunately it does not eliminate the occasional "code 19" halts in between.
Any ideas? Why is "code 19" (Pre-memory PCH initialization is started) so persistent here? Is there any reason to look at the PCH itself?
Cheers,
Willy
Btw, interesting observation that may be indicating some related condition:
Disabled CPU spread spectrum - R4E got right into "code 19" and wanted me to set DDR3-1333 mode.
Resetting the CPU spread spectrum back to Auto resulted in a normal boot and no problems with the DDR3-1600 setting!
How about that! (Normally, all OC gurus would tell you to disable the CPU & PCIe spread spectrum - here it seems to be working the other way around...)
Status: Almost perfect now, occasional "code 19" halts (rarely) - not due to the workload, something else weird is going on.
It seems to me that after a "code 19" halt, even though the system has been reset, the R4E is picking up previous "code 19" trail on startup and doesn't really start fresh.
This leads to repeating boot (POST) code 19 problems, even though when the system once has started it seems to run rock solid now (e.g. full load LinX without any sign of issues etc.).
Actions taken thus far:
WAVE 1
1. Started with a "clean" table: Reloaded default BIOS (F5) -> save & reset (F10). Note that the CPU is overclocked by default! (target 3900G)
2. Disabled all unused components/features (e.g. BT, Asmedia OPROM, etc.) -> save & reset
3. Set OC mode to manual and typed in the Kingston XMP values manually (VCCSA left to Auto!) -> save & reset
Gave it one pass with Memtest86+ (150 minutes) = 0 errors (Memtest86+ version 4.20)
On exit from Memtest86+ got "code 19" halt
Restarted the rig and it stopped on BIOS setup with a message: "Overclocking failed..."
Observed that the BIOS was pushing DDR3-1333 mode, rather than the manually set target DDR3-1600 (XMP)
WAVE 2
1. Increased the VCCSA voltage stepwise (0.05v increments) to 1.05v -> save & reset several times = all Ok
2. Started Windows and ran various stress tests (AIDA64, SiSandra, LinX), repeatedly - everything Ok
3. At some point Windows crashed on starting AIDA64 (when loading BIOS data) -> rebooting halted on "code 19" again
This time R4E went into a real "code 19" cycle, and it was very hard to get out of it. It seems that crashing Windows is causing harder R4E hangups.
Eventually the boot moved on and halted on code 94, and then finally next reboot went back to a working system.
I was curious so without changing anything else (other than the VCCSA=1.05v adjustment), I gave this setup another pass with Memtest86+ = 0 errors!
Basically, Memtest86+ has never complained about the memory or the way it works. Something else is going on...
Indeed, on exiting Memtest86+ I got the "code 19" halt again.
However, rebooting the system went fine this time, and I guess R4E just don't like the way Memtest86+ is terminating
Note that I've been running all the way with memory in a DDR3-1600 mode (checked regularly with CPU-Z)
WAVE 3
I decided to test the rig with a heavy workload, just to see if stressing the system is what is causing these problems, or there's something else.
1. Started with running AIDA64 stability stresstest for 30 minutes - no problems and everythings was Ok during
2. Ran full load LinX 0.6.4 for 30 minutes - no problems whatsoever (I was actually able to run other programs on top, the system didn't get sluggish at all!)
3. I went thru all the SiSandra queries collecting various HW information - worked like a charm
So at this point, I'm convinced that the RAM in itself is Ok, once started it's working fine, while "code 19" halts appear (now only occasionally) in various rebooting scenarios.
Finally, I entered 1.15v for VCCSA and 10 for CAS and tRCD as recommended, but unfortunately it does not eliminate the occasional "code 19" halts in between.
Any ideas? Why is "code 19" (Pre-memory PCH initialization is started) so persistent here? Is there any reason to look at the PCH itself?
Cheers,
Willy
Btw, interesting observation that may be indicating some related condition:
Disabled CPU spread spectrum - R4E got right into "code 19" and wanted me to set DDR3-1333 mode.
Resetting the CPU spread spectrum back to Auto resulted in a normal boot and no problems with the DDR3-1600 setting!
How about that! (Normally, all OC gurus would tell you to disable the CPU & PCIe spread spectrum - here it seems to be working the other way around...)
Last edited: