Are these addresses safe to be blocked?

[DevBT]

Summary

Report Date	2021-08-27 10:03:52
Generated by	MemTest86 V9.2 Free (64-bit) Visit MemTest86.com to Upgrade to Pro
Result	FAIL

System Information

EFI Specifications	2.70
System
Manufacturer	Micro-Star International Co., Ltd.
Product Name	MS-7B89
Version	1.0
Serial Number	To be filled by O.E.M.
BIOS
Vendor	American Megatrends International, LLC.
Version	1.H3
Release Date	04/20/2021
Baseboard
Manufacturer	Micro-Star International Co., Ltd.
Product Name	B450M MORTAR (MS-7B89)
Version	1.0
Serial Number	I816011555
CPU Type	AMD Ryzen 5 2600X Six-Core
CPU Clock	3340 MHz [Turbo: 4150.4 MHz]
# Logical Processors	12 (6 enabled for testing)
L1 Cache	12 x 96K (73784 MB/s)
L2 Cache	12 x 512K (53761 MB/s)
L3 Cache	1 x 16384K (27608 MB/s)
Memory	32710M (14761 MB/s)
Number of RAM SPDs detected	4
SPD #0	8GB DDR4 XMP PC4-25600
	G Skill Intl / F4-3200C16-8GTZR
	16-18-18-38 / 3200 MHz / 1.350V
SPD #1	8GB DDR4 XMP PC4-25600
	G Skill Intl / F4-3200C16-8GTZR
	16-18-18-38 / 3200 MHz / 1.350V
SPD #2	8GB DDR4 XMP PC4-25600
	G Skill Intl / F4-3200C16-8GTZR
	16-18-18-38 / 3200 MHz / 1.350V
SPD #3	8GB DDR4 XMP PC4-25600
	G Skill Intl / F4-3200C16-8GTZR
	16-18-18-38 / 3200 MHz / 1.350V
Number of RAM slots	4
Number of RAM modules	4
DIMM Slot #0	8GB DDR4 PC4-17000
	Unknown / F4-3200C16-8GTZR / 00000000
	2133 MHz
DIMM Slot #1	8GB DDR4 PC4-17000
	Unknown / F4-3200C16-8GTZR / 00000000
	2133 MHz
DIMM Slot #2	8GB DDR4 PC4-17000
	Unknown / F4-3200C16-8GTZR / 00000000
	2133 MHz
DIMM Slot #3	8GB DDR4 PC4-17000
	Unknown / F4-3200C16-8GTZR / 00000000
	2133 MHz

Result summary

Test Start Time	2021-08-27 01:51:33
Elapsed Time	7:28:52
Memory Range Tested	0x0 - 840000000 (33792MB)
CPU Selection Mode	Parallel (All CPUs)
CPU Temperature Min/Max/Ave	45C/63C/52C
RAM Temperature Min/Max/Ave	41C/47C/44C
# Tests Passed	45/48 (93%)

Lowest Error Address	0x440371140 (17411MB)
Highest Error Address	0x782F432C8 (30767MB)
Bits in Error Mask	0000000000000680
Bits in Error	3
Max Contiguous Errors	1
CPUs that detected memory errors	{ 2, 6, 10 }

Bits in Error

63

0

Test	# Tests Passed	Errors
Test 0 [Address test, walking ones, 1 CPU]	4/4 (100%)	0
Test 1 [Address test, own address, 1 CPU]	4/4 (100%)	0
Test 2 [Address test, own address]	4/4 (100%)	0
Test 3 [Moving inversions, ones & zeroes]	4/4 (100%)	0
Test 4 [Moving inversions, 8-bit pattern]	4/4 (100%)	0
Test 5 [Moving inversions, random pattern]	4/4 (100%)	0
Test 6 [Block move, 64-byte blocks]	4/4 (100%)	0
Test 7 [Moving inversions, 32-bit pattern]	4/4 (100%)	0
Test 8 [Random number sequence]	4/4 (100%)	0
Test 9 [Modulo 20, ones & zeros]	1/4 (25%)	20
Test 10 [Bit fade test, 2 patterns, 1 CPU]	4/4 (100%)	0
Test 13 [Hammer test]	4/4 (100%)	0

Last 10 Errors

2021-08-27 08:17:28 - [Data Error] Test: 9, CPU: 6, Address: 782F432C8, Expected: 85BF5FA1, Actual: 85BF5F21

2021-08-27 08:10:22 - [Data Error] Test: 9, CPU: 2, Address: 440371140, Expected: 9DC35EB8, Actual: 9DC35CB8

2021-08-27 08:10:21 - [Data Error] Test: 9, CPU: 2, Address: 440374E10, Expected: 64E75EA6, Actual: 64E75CA6

2021-08-27 08:10:20 - [Data Error] Test: 9, CPU: 2, Address: 440377B88, Expected: 64E75EA6, Actual: 64E75AA6

2021-08-27 08:10:20 - [Data Error] Test: 9, CPU: 2, Address: 440371140, Expected: 64E75EA6, Actual: 64E75CA6

2021-08-27 08:10:17 - [Data Error] Test: 9, CPU: 2, Address: 440374E10, Expected: 06446E31, Actual: 06446C31

2021-08-27 08:10:17 - [Data Error] Test: 9, CPU: 2, Address: 440377B88, Expected: 06446E31, Actual: 06446A31

2021-08-27 08:10:16 - [Data Error] Test: 9, CPU: 2, Address: 440371140, Expected: 06446E31, Actual: 06446C31

2021-08-27 06:08:43 - [Data Error] Test: 9, CPU: 10, Address: 440377B88, Expected: E1719E24, Actual: E1719A24

2021-08-27 06:08:43 - [Data Error] Test: 9, CPU: 10, Address: 440371140, Expected: E1719E24, Actual: E1719C24

As the title says, is it safe to block this addresses from Windows 10? I'm assuming that as long as their blocked, Windows won't use these, is that correct?

[David (PassMark)]

Lowest Error Address	0x440371140 (17411MB)
Highest Error Address	0x782F432C8 (30767MB)

This is a huge range of addresses. Almost 50% of your RAM. So it doesn't make to block them all.
Better to find the bad stick and replace it.

See also
https://www.memtest86.com/troubleshooting.htm

[DevBT]

I was actually thinking just blocking the actual addresses listed in the errors section. Is it not possible?

[David (PassMark)]

Most RAM has a lifetime warranty. So it is basically free to replace once you are sure it is bad. And the above report shows pretty conclusively that it is bad.

By default only the last 10 errors are listed in the MemTest86 report. So there might be other addresses. Also if it low quality RAM, it would not be surprising to find another group of bad addresses even if you block these.

Details about black listing addresses can be found here
https://www.memtest86.com/blacklist-...emorylist.html

In Linux it is fairly straight forward. In Windows there are a bunch of quirks which make it hard.

[DevBT]

I'll have to resort to blocking the addresses for now. The shop where I bought the modules only offer 1-year shop warranty and they are not in the list of "authorized G.SKILL retailer or reseller" and doing an RMA would be more of a hassle than buying a new one. I might just buy a new one, but not really that urgent.

One question though, I sort of removed the faulty modules but I forgot which module was from which slot. Will the addresses reported be the same if ever I returned them to the wrong slots compared when I did the test?

[DevBT]

I've decided to just run the tests again but just the test with the one with errors. So I'll get the address properly.

[David (PassMark)]

Will the addresses reported be the same if ever I returned them to the wrong slots compared when I did the test?

No. The addresses will change if the slot order was changed. Interleaving addresses in dual channel mode causes this.

[DevBT]

Figured, I just ran the test again and fortunately, I put them in the same slots.

I've blocked the faulty addresses now and seems to be working fine so far.