3700x Underperforming

[StarsMars]

I have a 3700x, under a nh-d15 on a Asus crosshair vii hero x470.

My best cinnebench r15 score was 2169 at some point a few months ago. Since then it hasn't gone much above 2050

Core clocks seem to be around 4.125ghz at best.

Running pbo.

I'm an overclocking noob.

 

Does this sound right?

[5x5]

Thermals and power? Use Ryzen master to track those accurately. Also, update the BIOs and firmware.

[StarsMars]

34 minutes ago, 5x5 said:

Thermals and power? Use Ryzen master to track those accurately. Also, update the BIOs and firmware.

When test/applying pbo profiles, it stays at 69c.

Running cpuid cpu stress test temps hit 72c

Updated the bios before posting.

 

 

 

 

^while running cpuid cpu stress

[StarsMars]

I know there was some drama around false boost ratings on the box. But is 4.1 normal? That's a bit off of 4.4ghz

[xg32]

3 minutes ago, StarsMars said:

I know there was some drama around false boost ratings on the box. But is 4.1 normal? That's a bit off of 4.4ghz

4.1 all core is normal, 4.4 all core for everyday use degrades the chip.

[StarsMars]

Just now, xg32 said:

4.1 all core is normal, no1 can oc to 4.4 without degrading the chip.

Well it's good to know I don't just have a bad chip.

So 4.4 is just a lie?

[nick name]

Where do you live?  Summer is approaching and if your ambient temp is higher now than you seeing a reduction now compared to the winter time makes sense.  

[StarsMars]

4 minutes ago, nick name said:

Where do you live?  Summer is approaching and if your ambient temp is higher now than you seeing a reduction now compared to the winter time makes sense.  

I'm in air conditioning. It's kept at 24c.

[nick name]

39 minutes ago, StarsMars said:

I'm in air conditioning. It's kept at 24c.

So it wasn't cooler in your place during the winter?

[nick name]

Also, when you completed your build I'm assuming you did a fresh install of Windows 10?  Perhaps now the system has more tasks running than before?

[StarsMars]

36 minutes ago, nick name said:

So it wasn't cooler in your place during the winter?

No, it's always the same in here.

21 minutes ago, nick name said:

Also, when you completed your build I'm assuming you did a fresh install of Windows 10?  Perhaps now the system has more tasks running than before?

No, I didn't do a fresh install. Maybe I should go for it.

[Triboal]

2 hours ago, StarsMars said:

Well it's good to know I don't just have a bad chip.

So 4.4 is just a lie?

No, 4.4 GHz is single core and if you run Cinebench R15/R20 single threaded workload you should see one of the cores hit closer to that.

[StarsMars]

I checked the windows power plan. I guess after some windows update, it turned off ryzen optimized power plan. It helped a bit. R15 score is now 2111

[maizenblue]

My 3700x boosts a bit higher. Ill look for you later after work but i think it gets up to 4.3, but thats not using pbo so its not all core, just a few cores or single core boosting. 
 

both amds and intels max boost seem to be hit or miss, with some boosting to advertised max speeds and some not. The reason might not necessarily be silicon lottery but a variety of factors. 

[Guest]

51 minutes ago, maizenblue said:

My 3700x boosts a bit higher. Ill look for you later after work but i think it gets up to 4.3, but thats not using pbo so its not all core, just a few cores or single core boosting. 
 

both amds and intels max boost seem to be hit or miss, with some boosting to advertised max speeds and some not. The reason might not necessarily be silicon lottery but a variety of factors. 

The SenseMi algorithm is based on temps. Lower the temps, higher average boostz. 

[StarsMars]

3 hours ago, maizenblue said:

My 3700x boosts a bit higher. Ill look for you later after work but i think it gets up to 4.3, but thats not using pbo so its not all core, just a few cores or single core boosting. 
 

both amds and intels max boost seem to be hit or miss, with some boosting to advertised max speeds and some not. The reason might not necessarily be silicon lottery but a variety of factors. 

 

2 hours ago, ShrimpBrime said:

The SenseMi algorithm is based on temps. Lower the temps, higher average boostz. 

Checked a bit more. Looks like it's running about 4,150 to 4,290 in games. I've seen it randomly hit 4.325 in desktop.

Guess I was confusing all core with single peak.

 

I still want to see if there are any settings in the bios I should change without manually overclocking.

Ram is 3600cl16 btw

[Guest]

43 minutes ago, StarsMars said:

 

Checked a bit more. Looks like it's running about 4,150 to 4,290 in games. I've seen it randomly hit 4.325 in desktop.

Guess I was confusing all core with single peak.

 

I still want to see if there are any settings in the bios I should change without manually overclocking.

Ram is 3600cl16 btw

Ive seen these running hot enough to derate to 3.9ghz. 

So you're doing just fine. 

Run it like it is, I see no need to change anything.

[StarsMars]

On 3/24/2020 at 11:04 AM, ShrimpBrime said:

Ive seen these running hot enough to derate to 3.9ghz. 

So you're doing just fine. 

Run it like it is, I see no need to change anything.

Good to hear, thanks for the help guys

[maizenblue]

On 3/24/2020 at 7:41 AM, ShrimpBrime said:

The SenseMi algorithm is based on temps. Lower the temps, higher average boostz. 

How low do the temps need to be to meet max boost? Running cinebench r20 and a max temp of 65 using a d15s it still doesnt quite boost to 4.4ghz.

[Guest]

2 hours ago, maizenblue said:

How low do the temps need to be to meet max boost? Running cinebench r20 and a max temp of 65 using a d15s it still doesnt quite boost to 4.4ghz.

Well should be close to max boost about 50mhz below max single core, give or take. 

You have good testimate to validate higher boost clocks with lower temps.

 

My 2700x with 8c water delta boost all cores 4300mhz on auto, no PBO oc needed. This was also advertised as a feature when AMD first gen was released. 

 

The issue with 3000 series is the FIT contrler limits the voltage. Around 1.320v (give or take) to prevent too high current draw at the socket.

 

And thus we always see the inquire if 1.4v is safe. Yes and no. 1.4v safe on low current pull ie: single core boosts.