The 2016 FX overclocking results thread

[zentan]

It is running at 5.1, and I just proved it isn't throttling. It just doesn't score as well as my regular 8370.

There's absolutely no reason that it would have such a big difference in score,after all it's the same architecture with same number of cores.It probably is throttling or not sustaining the clocks well and hence The Stilt is right.It just isn't running at 5.1 for the whole period.
Also the 55C temperature at 5.1GHz with Wraith just can't be of a sustained clock,no point in reporting such temperatures.

 

[kwikgta]

There's absolutely no reason that it would have such a big difference in score,after all it's the same architecture with same number of cores.It probably is throttling or not sustaining the clocks well and hence The Stilt is right.It just isn't running at 5.1 for the whole period.
Also the 55C temperature at 5.1GHz with Wraith just can't be of a sustained clock,no point in reporting such temperatures.

I've been busy.....I ended up testing both of my 8370's and the 8370e in the test rig on the 990fx gaming board and the "e" scored considerably lower, then I tested all 3 chips in the main rig on my UD5 board, with the same results. I even tried running them all at 4.3 with the exact same bios settings. and no matter what, the "e" scores way less on every bench mark I ran. I have no idea where the difference is, but my e chip is the lesser chip of the 3 for sure.

 

[Abwx]

Hardware.fr got their FX8370E at 4.8GHz at 1.428V under prime 95, they explicitely stated that at 1.45V they could do a few bench at 5.1GHz.

http://www.hardware.fr/focus/99/amd-fx-8370e-fx-8-coeurs-95-watts-test.html

So what is the opinion of the skeptics by here knowing that Hfr chip was just slightly better (barely 2% lower voltage at 4.4) than the FX8320E that AT got for their review..?

 

[myocardia]

It's worth mentioning that I'm using a Fluidyne automotive radiator with 4x200mm fans in push/pull, too.

A radiator!? You do mean a heater core, don't you?

 

[Madpacket]

If the scores are different you're throttling on the 8370e simple as that. Try upping the voltage on the 8370e and I bet your scores will go up.

 

[zentan]

Hardware.fr got their FX8370E at 4.8GHz at 1.428V under prime 95, they explicitely stated that at 1.45V they could do a few bench at 5.1GHz.

http://www.hardware.fr/focus/99/amd-fx-8370e-fx-8-coeurs-95-watts-test.html

So what is the opinion of the skeptics by here knowing that Hfr chip was just slightly better (barely 2% lower voltage at 4.4) than the FX8320E that AT got for their review..?

Did Hardware.fr say 5.1GHz w/ Wraith cooler @1.425V at 55C while stress testing the cpu and got a score corresponding to that of ~4.7GHz?
You don't need to make A as B,those were significantly different.

 

[kwikgta]

I also tried all 3 of my cpu's at 4300 MHz and stock voltage in 2 different motherboards and my "e" chip still scored way less. Why? Hell if I know, but that's why it is in my teenage sons rig now instead of one of mine.

 

[myocardia]

I also tried all 3 of my cpu's at 4300 MHz and stock voltage in 2 different motherboards and my "e" chip still scored way less. Why? Hell if I know, but that's why it is in my teenage sons rig now instead of one of mine.

I've read reports from forum posters over the past 18 months of just that type of difference between the 8370 and the 8370e. Did you replace your son's motherboard, when you made the switch? While reading your posts in this thread, I was getting the feeling that maybe a portion of the 8370e's problem was the ancient 970 chipset in its motherboard. Whatcha think about that?

edit: Or did your son's motherboard already have a 990 chipset?

 

[MongGrel]

NM, I won't bother, for several reasons.

 

[SlowSpyder]

A radiator!? You do mean a heater core, don't you?

A heat exchanger, it's probably made for a transmission or oil cooler, something along those lines... it's about 18"x9". I suppose 'radiator' isn't really the correct term.

 

[myocardia]

NM, I won't bother, for several reasons.

I've been busy.....I ended up testing both of my 8370's and the 8370e in the test rig on the 990fx gaming board and the "e" scored considerably lower, then I tested all 3 chips in the main rig on my UD5 board, with the same results. I even tried running them all at 4.3 with the exact same bios settings. and no matter what, the "e" scores way less on every bench mark I ran. I have no idea where the difference is, but my e chip is the lesser chip of the 3 for sure.

Yes, I had missed this post. That was the last question in my mind, and this is the first post I've read on any forum where an 8370e has been benchmarked in one (or two in this case) higher-end motherboards, that have the best AM3+ chipset that AMD makes, against other non-e FXs.

 

[myocardia]

A heat exchanger, it's probably made for a transmission or oil cooler, something along those lines... it's about 18"x9". I suppose 'radiator' isn't really the correct term.

That's too funny, Steve. I actually came very, very close to editing that post last night, to add transmission cooler to the mix. I've never heard of anyone using a transmission or an oil cooler for CPUs, since heater cores are about exactly the same size, except the heater cores are about twice as thick, which in my mind would give better cooling for the same volume of airflow, but anything is definitely possible.

Hmm, actually Fluidyne's heat exchangers are being used for street racing (small-time, not Street Outlaws-class) intercooler upgrades for factory turbo or supercharged cars. Both Jegs and Summit Racing are carrying them now. Also, Fluidyne does not make heater cores, only transmission & oil coolers, along with the heat exchangers, like you have.

 

[SlowSpyder]

That's too funny, Steve. I actually came very, very close to editing that post last night, to add transmission cooler to the mix. I've never heard of anyone using a transmission or an oil cooler for CPUs, since heater cores are about exactly the same size, except the heater cores are about twice as thick, which in my mind would give better cooling for the same volume of airflow, but anything is definitely possible.

Hmm, actually Fluidyne's heat exchangers are being used for street racing (small-time, not Street Outlaws-class) intercooler upgrades for factory turbo or supercharged cars. Both Jegs and Summit Racing are carrying them now. Also, Fluidyne does not make heater cores, only transmission & oil coolers, along with the heat exchangers, like you have.

Yup, it was dirt cheap compared to an actual PC radiator anywhere close to this size. Here is the one I have with what looks to be an aftermarket RX7 radiator to give you an idea of its size. It's thin (about 1" thick), but lots of surface area and with 200mm fans in push/pull it seems to keep up well enough for my use. It was something worth trying.

 

[kwikgta]

I've read reports from forum posters over the past 18 months of just that type of difference between the 8370 and the 8370e. Did you replace your son's motherboard, when you made the switch? While reading your posts in this thread, I was getting the feeling that maybe a portion of the 8370e's problem was the ancient 970 chipset in its motherboard. Whatcha think about that?

edit: Or did your son's motherboard already have a 990 chipset?

All 3 chips were tested in my 990FXa - Gaming board and my UD5 board at overclocked speeds and at 4.3 speeds. The "e" chip lost big time to my desktop chips in both boards and all speeds tested. I then installed it into the son's pc on a 990-UD3 board and set it at 4.3 since he was only running an old 4170 quad core. I didn't bench it on the UD3 board but he's happy with it.

 

[SlowSpyder]

They should score the same, in theory. My guess is something was different, HT speed, ram speed, turbo or c-state on / off on one and not the other. But who knows, probably not that important if you're content with how everything is running. Sounds like you got some good silicon. My 2016 FX8320e on the other hand... not such a strong showing.

 

[kwikgta]

The only thing that I'm questioning in my mind now is that maybe my "e" chip was some kind of refurb chip. I bought it on fleabay for $150 when everyone else had them around $200 and up. The seal was cut on the box when I got it. Guess I will never know.

 

[myocardia]

All 3 chips were tested in my 990FXa - Gaming board and my UD5 board at overclocked speeds and at 4.3 speeds. The "e" chip lost big time to my desktop chips in both boards and all speeds tested. I then installed it into the son's pc on a 990-UD3 board and set it at 4.3 since he was only running an old 4170 quad core. I didn't bench it on the UD3 board but he's happy with it.

Yeah, sorry...I had somehow missed/skipped over your longer post stating that you had benchmarked it in your 990FX boards. And yeah, I'd take a stock speed 4170 over a stock speed 8370, for gaming, but I'd take a 4.3 Ghz on all cores 8370e over any 4170, at any speed.

 

[kwikgta]

Yeah, sorry...I had somehow missed/skipped over your longer post stating that you had benchmarked it in your 990FX boards. And yeah, I'd take a stock speed 4170 over a stock speed 8370, for gaming, but I'd take a 4.3 Ghz on all cores 8370e over any 4170, at any speed.

No problem. Im just glad I got my best chip of 3 in the main rig now.

 

[Madpacket]

If performance is different at the same settings you should try to find out why. There's obviously something different with the E CPU so it would be ideal to diagnose why this happening. I still suspect it's a voltage issue but you would need to do some serious monitoring to understand the differences.

 

[Abwx]

Did Hardware.fr say 5.1GHz w/ Wraith cooler @1.425V at 55C while stress testing the cpu and got a score corresponding to that of ~4.7GHz?
You don't need to make A as B,those were significantly different.

Dont know the scores but they didnt mention that there was any issue, they use a good cooler but nothing exceptional if the 4790K overclocking tests are of any indication, and the Wraith cooler is said to be quite efficient for a stock cooler.

http://www.hardware.fr/focus/98/core-i7-4790k-test-devil-s-canyon-debarque.html

For the 8370E they restricted the voltage to 1.45V and didnt test with a higher value.

. My 2016 FX8320e on the other hand... not such a strong showing.

AT got a quite good sample but it still displayed a difference with the 8370E, this latter is obviously better once frequency get over 4.3 or so, Computerbase.de 8370E sample also reached 5GHz.

 

[rancherlee]

Cooling the chip isn't an issue @ 1.45v, its keeping the VRM's from overheating and throttling the chip. My week 17 14' 8320 would prime 4.3 @ 1.3v with the crappy cooler it came with under prime for hours and stay just under the thermal limits. I scored an H90 CLC cheap and bumped it to 4.5@ default 1.35v and it stays under 55* BUT after 30 minutes of Prime the VRM's would get too hot and throttle the chip. I've added a fan blowing on the backside of the socket/motherboard and that stopped the throttling BUT I haven't tried pushing further. This is with a 6+2 board, going for 5ghz @ 1.45v needs an 8+2 board to avoid throttling.

 

[SlowSpyder]

AT got a quite good sample but it still displayed a difference with the 8370E, this latter is obviously better once frequency get over 4.3 or so, Computerbase.de 8370E sample also reached 5GHz.

Yea, I cheaped out and figured I'd give the FX8320e a try. For the record it still overclocks ~40%, nothing to sneeze at. It's just that my FX9370 goes a fair bit higher. Had my FX9370 not set expectations unrealistically high, I would have been perfectly happy with this FX8320e.

 

[Vortex6700]

Cpu: 9590 5.1ghz (was 5.5)
Voltage: 1.45 (load) 1.5 (idle)
Cooler: Antec 240mm sealed w 4x 120mm push/pull
Motherboard: Asus Crosshair V Formula
Purchased: 2014

 

[FlanK3r]

I tested few days ago FX-8370s. One older 2014 batch and one new 2015. New one was crappy at one modul, hard to get 4640 MHz with 1.45V stable. Old one with the same voltage 4916 MHz stable.

 

[SlowSpyder]

Cpu: 9590 5.1ghz (was 5.5)
Voltage: 1.45 (load) 1.5 (idle)
Cooler: Antec 240mm sealed w 4x 120mm push/pull
Motherboard: Asus Crosshair V Formula
Purchased: 2014

Was 5.5GHz?? I was able to get into Windows @ 5.5GHz with only one module enabled. Stability wasn't very good. That'd be a pretty good achievement if you could run that high! My FX9370 ran 5.34GHz on all four modules for a while... outside of super heavy loads (AVX x8 threads) it was fairly stable, not perfect though.