Asustor Lockerstor 4 (AS6604T) NAS Review
Hitman III: PC graphics perf benchmark review
TeamGroup CX2 1TB SATA3 SSD review
EVGA GeForce RTX 3070 FTW3 Ultra review
Corsair 5000D PC Chassis Review
NZXT Kraken X63 RGB Review
ASUS Radeon RX 6900 XT STRIX OC LC Review
TerraMaster F5-221 NAS Review
MSI Radeon RX 6800 XT Gaming X TRIO Review
Sapphire Radeon RX 6800 NITRO+ review
Ryzen 5600X and 5800X CPUs Can Have Two CPU dies
Or at least some of them. Of course, 6 and 8-core processors are expected to have just one die (1x CCD). However, we do not know how AMD is binning products. In this case, the second die might have an issue and is disabled or in deep sleep mode.
Yuri (1usmus) noticed that there are models out there that have a second CCD on board. While working on his Clock Tuner for Ryzen tool, the Ryzen 5 5600X was testing turned out to have a second die. This second CCD could also be read but not used. The second chiplet appears to be permanently in Deep Sleep Mode Occasionally; something seemed to happen with it because the clock speed read increased to 550 MHz.
Presumably, the Ryzen 5 proc was originally a Ryzen 9 5900X, of which one of the two 6-core chiplets did not work properly and got binned to a 5600X. Given the still somewhat scarce availability of the Ryzen 5000 CPUs, it is not surprising that AMD still sells such a chip. Interesting is that the second die does not seem to be disabled or laser cut. There may be a way to unlock the remaining cores, similar to what one could do with certain Phenom II processors, for example. CTR will have support for detecting 'dormant' CCDs. It likely is released around the end of this month.
H83
Senior Member
Posts: 3344
Joined: 2009-09-08
Senior Member
Posts: 3344
Joined: 2009-09-08
#5876101 Posted on: 01/11/2021 03:26 PM
Those versions with two dies should have a different naming from the other ones, to identify them, and should be cheaper....
Those versions with two dies should have a different naming from the other ones, to identify them, and should be cheaper....
Richard Nutman
Senior Member
Posts: 194
Joined: 2018-08-30
Senior Member
Posts: 194
Joined: 2018-08-30
#5876108 Posted on: 01/11/2021 03:44 PM
I don't think the main 6 or 8 cores are spread over 2 dies. So it's not like you're getting a worse chip compared to others with higher latency on some cores. It's just a failed dual die chip, the other die is completely inactive.
Those versions with two dies should have a different naming from the other ones, to identify them, and should be cheaper....
I don't think the main 6 or 8 cores are spread over 2 dies. So it's not like you're getting a worse chip compared to others with higher latency on some cores. It's just a failed dual die chip, the other die is completely inactive.
asturur
Senior Member
Posts: 975
Joined: 2010-05-12
Senior Member
Posts: 975
Joined: 2010-05-12
#5876109 Posted on: 01/11/2021 03:44 PM
The only possible reason I can come up with for these dies to exist is to maybe share something that isn't core-specific, like the L3 cache. I'm sure there are situations where all cores are good but maybe L3 isn't. Otherwise, it makes no sense why some would have this while others don't. It's not like Threadripper where they were just trying to distribute the load of the IHS using dummy chiplets.
It'd be interesting if there are differences in benchmarks.
Not sure how l3 cache outside the chiplet is usefull tho.
Wouldn't it be slow?
Important is that you do not get 8 or 6 cores split 4 +4 or 3+3 for some other weird reasons
The only possible reason I can come up with for these dies to exist is to maybe share something that isn't core-specific, like the L3 cache. I'm sure there are situations where all cores are good but maybe L3 isn't. Otherwise, it makes no sense why some would have this while others don't. It's not like Threadripper where they were just trying to distribute the load of the IHS using dummy chiplets.
It'd be interesting if there are differences in benchmarks.
Not sure how l3 cache outside the chiplet is usefull tho.
Wouldn't it be slow?
Important is that you do not get 8 or 6 cores split 4 +4 or 3+3 for some other weird reasons
Silva
Senior Member
Posts: 1360
Joined: 2013-06-04
Senior Member
Posts: 1360
Joined: 2013-06-04
#5876111 Posted on: 01/11/2021 03:51 PM
If all of them meet the 5600X standard, I don't know why that would be.
As the article states, probably a 5900X that failed and one of the chiplets was disabled.
Those versions with two dies should have a different naming from the other ones, to identify them, and should be cheaper....
If all of them meet the 5600X standard, I don't know why that would be.
As the article states, probably a 5900X that failed and one of the chiplets was disabled.
Click here to post a comment for this news story on the message forum.
Senior Member
Posts: 5591
Joined: 2012-11-10
The only possible reason I can come up with for these dies to exist is to maybe share something that isn't core-specific, like the L3 cache. I'm sure there are situations where all cores are good but maybe L3 isn't. Otherwise, it makes no sense why some would have this while others don't. It's not like Threadripper where they were just trying to distribute the load of the IHS using dummy chiplets.
It'd be interesting if there are differences in benchmarks.