I bought an FX-8350 as a drop-in replacement to my 1090T just a little over 2 months ago (July 1).
It died about a week ago, making it last almost exactly 2 months.
I'm sure it is the CPU because I am typing this on the exact same PC, just with the old 1090T swapped back.
Short story:
With a really crappy thermal paste, it died with a very modest OC (4.3GHz, load vcore of 1.396V, HPC Mode on, CnQ on) while running Prime95 for about 2 hours. Temps reached 74-76C. It died and wouldn't post anymore. A CPU swap to the 1090T confirmed the death of the 8350. 1 week later (today), I tried to swap it back in to just verify 100% that it is indeed dead. (Fear not, brothers, it did not really die - it just turned into an expensive keychain)
I'm not sure why thermal throttling didn't take place. CnQ was even on. Perhaps it was HPC Mode being on. Yeah, it's probably that. When it is on, 8-core Turbo is the same as the single-core Turbo, and during Prime tests it removes the occasional clock-drop to 3GHz that lasts for <1 second. Had no idea it completely removes any thermal protection (if it is indeed the culprit, I'm just guessing).
So there you go, a mild voltage bump + temps well into mid-70C killed it when ran for 2-3 hours straight at max load.
Long story:
It was cheap (<$200) and was a drop-in replacement. I had a 1090T rig that was perfectly fine, but I decided Piledriver wasn't as bad as Bulldozer and was worth getting as a drop-in replacement (not as a new rig, though).
It arrived July 1, then spent 2 weeks on the shelf because I was busy, then finally got the time to bench my 1090T before swapping it out and then benching the 8350 and comparing. (I know I mentioned some of these results already, how a stock 8350 was already a bit better than a 3.7GHz 1090T).
After 3 weeks, I wanted to try something on the 1090T so I swapped it right back. By the time I was ready to swap back the 8350, I realized I was out of thermal paste (Tuniq TX-2, a few tubes of which served me well for the past 4-5 years).
I went down my favorite store and was shocked to learn that they no longer had Tuniq in stock (no TX-2, and none of the newer TX-3 and TX-4). Instead, they had an old Cooler Master and a Thermaltake TG-2. Both are crap, the Cooler Master because the stocks are so old it's way past its shelf life (they stopped manufacturing this model for a few years now). The Thermaltake TG-2 is absolutely hands-down the worst thermal paste you could get your hands-on (it doesn't even beat the expired Cooler Master).
Anyway, having no choice, I bought both and wished myself luck. Temps were much much higher (I don't mean 4-5C, I mean >10C). Ah, the TG-2 really does not disappoint - it doesn't just fail, it fails spectacularly! I've seen two round-ups that had it at the very bottom of the chart. I can certainly confirm.
Anyway, I tried different application methods in an attempt to make do with what I have until I could get a Zalman STG-1/2 (the only other SKU the store carries, but was out of stock at the moment I bought).
The motherboard only allows a modest 4.3GHz OC on the 8350 due to lack of higher vcore options and lack of vdroop control / LLC. At 4.3GHz, max load vcore I can get is ~ 1.396V, just a bit shy of 1.4V.
Using a Thermaltake Frio OCK at max fan with the Tuniq, I'd be getting max load temps (Prime95 Large FFT) of ~57-59, HPC Mode on. I tried that same profile on the TG-2, and temps soared to over 70C, settling at around 74-76C.
Here's what I was thinking. AMD doesn't publish max temps AND vcore ranges for this part. If I go by the FX-8150 (direct ancestor), max temp should only be 61C (but at what range of operating voltages? AMD doesn't publish). But Trinity mobile SKU's (like the A4-4300m I benched in VC&G) sporting the same Piledriver cores have a much higher limit (sites like CPUBoss and CPU World specify 100C, despite a vcore range of up to 1.35V). So, the same old Piledriver cores can take 1.35V and 100C? Well, if so, I'm still relatively ok (1.396V, 76C).
So, I just let it run continuously as I continued working on my laptop.
It just black-screened well into the 2nd hour, close to 3rd, probably.
Final thoughts:
So there you go. I'm not sure what exactly is helpful here, but off the top of my head I can name the ff:
1.) Really never settle on Thermaltake TG-2. TG-1 is really way better, or TG-2 is really just way way worse.
2.) Even at relatively low vcore (relative to most enthusiast overclocks of the FX-8350) of 1.396V, it's possible to kill an 8350 with temps up to 76C for a relatively moderate duration of 2-3 hours.
3.) HPC Mode (if it is indeed what stopped the CPU from throttling) is a dangerous weapon. It is very useful in that it allows better MT performance, but at the cost of shutting down all thermal protection.
4.) This is the very first CPU I've killed from both the Intel and AMD camp. I can remember killing a lot of lesser motherboards (various less expensive MSI and Gigabyte boards) due to overclocking, but never a CPU before. So this was pretty fun in a "well, that hasn't happened before" kind of way.
It died about a week ago, making it last almost exactly 2 months.
I'm sure it is the CPU because I am typing this on the exact same PC, just with the old 1090T swapped back.
Short story:
With a really crappy thermal paste, it died with a very modest OC (4.3GHz, load vcore of 1.396V, HPC Mode on, CnQ on) while running Prime95 for about 2 hours. Temps reached 74-76C. It died and wouldn't post anymore. A CPU swap to the 1090T confirmed the death of the 8350. 1 week later (today), I tried to swap it back in to just verify 100% that it is indeed dead. (Fear not, brothers, it did not really die - it just turned into an expensive keychain)
I'm not sure why thermal throttling didn't take place. CnQ was even on. Perhaps it was HPC Mode being on. Yeah, it's probably that. When it is on, 8-core Turbo is the same as the single-core Turbo, and during Prime tests it removes the occasional clock-drop to 3GHz that lasts for <1 second. Had no idea it completely removes any thermal protection (if it is indeed the culprit, I'm just guessing).
So there you go, a mild voltage bump + temps well into mid-70C killed it when ran for 2-3 hours straight at max load.
Long story:
It was cheap (<$200) and was a drop-in replacement. I had a 1090T rig that was perfectly fine, but I decided Piledriver wasn't as bad as Bulldozer and was worth getting as a drop-in replacement (not as a new rig, though).
It arrived July 1, then spent 2 weeks on the shelf because I was busy, then finally got the time to bench my 1090T before swapping it out and then benching the 8350 and comparing. (I know I mentioned some of these results already, how a stock 8350 was already a bit better than a 3.7GHz 1090T).
After 3 weeks, I wanted to try something on the 1090T so I swapped it right back. By the time I was ready to swap back the 8350, I realized I was out of thermal paste (Tuniq TX-2, a few tubes of which served me well for the past 4-5 years).
I went down my favorite store and was shocked to learn that they no longer had Tuniq in stock (no TX-2, and none of the newer TX-3 and TX-4). Instead, they had an old Cooler Master and a Thermaltake TG-2. Both are crap, the Cooler Master because the stocks are so old it's way past its shelf life (they stopped manufacturing this model for a few years now). The Thermaltake TG-2 is absolutely hands-down the worst thermal paste you could get your hands-on (it doesn't even beat the expired Cooler Master).
Anyway, having no choice, I bought both and wished myself luck. Temps were much much higher (I don't mean 4-5C, I mean >10C). Ah, the TG-2 really does not disappoint - it doesn't just fail, it fails spectacularly! I've seen two round-ups that had it at the very bottom of the chart. I can certainly confirm.
Anyway, I tried different application methods in an attempt to make do with what I have until I could get a Zalman STG-1/2 (the only other SKU the store carries, but was out of stock at the moment I bought).
The motherboard only allows a modest 4.3GHz OC on the 8350 due to lack of higher vcore options and lack of vdroop control / LLC. At 4.3GHz, max load vcore I can get is ~ 1.396V, just a bit shy of 1.4V.
Using a Thermaltake Frio OCK at max fan with the Tuniq, I'd be getting max load temps (Prime95 Large FFT) of ~57-59, HPC Mode on. I tried that same profile on the TG-2, and temps soared to over 70C, settling at around 74-76C.
Here's what I was thinking. AMD doesn't publish max temps AND vcore ranges for this part. If I go by the FX-8150 (direct ancestor), max temp should only be 61C (but at what range of operating voltages? AMD doesn't publish). But Trinity mobile SKU's (like the A4-4300m I benched in VC&G) sporting the same Piledriver cores have a much higher limit (sites like CPUBoss and CPU World specify 100C, despite a vcore range of up to 1.35V). So, the same old Piledriver cores can take 1.35V and 100C? Well, if so, I'm still relatively ok (1.396V, 76C).
So, I just let it run continuously as I continued working on my laptop.
It just black-screened well into the 2nd hour, close to 3rd, probably.
Final thoughts:
So there you go. I'm not sure what exactly is helpful here, but off the top of my head I can name the ff:
1.) Really never settle on Thermaltake TG-2. TG-1 is really way better, or TG-2 is really just way way worse.
2.) Even at relatively low vcore (relative to most enthusiast overclocks of the FX-8350) of 1.396V, it's possible to kill an 8350 with temps up to 76C for a relatively moderate duration of 2-3 hours.
3.) HPC Mode (if it is indeed what stopped the CPU from throttling) is a dangerous weapon. It is very useful in that it allows better MT performance, but at the cost of shutting down all thermal protection.
4.) This is the very first CPU I've killed from both the Intel and AMD camp. I can remember killing a lot of lesser motherboards (various less expensive MSI and Gigabyte boards) due to overclocking, but never a CPU before. So this was pretty fun in a "well, that hasn't happened before" kind of way.
Last edited: