R5 5600 performing worse than my old 3600

[Cantaloupeman]

So I started with the 3600 and decided to upgrade about a week or so after buying it. I want to be able to run BF2042 which is already a CPU demanding game. 

 

The 5600 seemed to have higher temps than I remembered the 3600 having so I reapplied thermal paste and the cooler a few times. Eventually I decided to put the old 3600 back in before returning it. The 5600 had lower temps in cinnebench by about 10 degrees, but my idle temps are slightly lower with the 3600 and my temps in BF2042 are about 6-7 degrees cooler. I also have less of an issue with stuttering with the 3600. My FPS is better all around. 

 

I noticed my core clock gets up to 4.0 GHZ with the 3600 but the 5600 only gets to 3.7. It has a lower TDP and and is using 100% of it. The 3600 is using only about 90% of the TDP so I'm wondering if this is why the core clocks are getting higher. 

I couldn't figure out how to enable PBO on either processor. Is this likely my issue or is there something else I'm not considering? 

[jaslion]

Rest of the specs?

 

Temps?

[Cantaloupeman]

11 minutes ago, jaslion said:

Rest of the specs?

 

Temps?

Gigabyte a520i Mobo

Wraith Spire (both processors)

Corsair TM30 Thermal Paste

Corsair SF750 PSU

XFX rx6600

Sandisk SATA SSD

Silverstone Sg13 Mini ITX Case

Generic old TT 120mm fan (Soon to be upgraded)

 

Temps for 5600 stock fan curve:

Idle- 42 Degrees

Cinnebench- 72 Degrees 1.05V 3.7Ghz 100% of 76W TDP

Bf2042- 76 Degrees

BF1- 77 Degrees

 

Temps for 3600 Stock fan curve:

Idle- 42-43 Degrees

Cinnebench- 84 Degrees 1.33V 3.9Ghz 91% of 88W TDP

Bf2042- 70 Degrees

 

 

 

 

 

[Cantaloupeman]

So I started with the 3600 and decided to upgrade about a week or so after buying it. I want to be able to run BF2042 which is already a CPU demanding game. 

 

The 5600 seemed to have higher temps than I remembered the 3600 having so I reapplied thermal paste and the cooler a few times. Eventually I decided to put the old 3600 back in to make sure I wasn't imagining things. The 5600 had lower temps in cinnebench by about 10 degrees, but my temps in BF2042 are about 6-7 degrees hotter with the 5600 and I get much more stuttering and lower frames.

 

I noticed my core clock gets up to 4.0 GHZ with the 3600 but the 5600 only gets to 3.7 in cinnebench. The 5600 has a lower TDP and and is using 100% of it. The 3600 is using only about 90% of the TDP so I'm wondering if this is why the core clocks are getting higher and thus giving me better frames and less stuttering.

 

I couldn't figure out how to enable PBO on either processor. Is this likely my issue or is there something else I'm not considering?

 

Here are my specs and the temps I recorded:

Gigabyte a520i Mobo

Wraith Spire (both processors)

Corsair TM30 Thermal Paste

Corsair SF750 PSU

XFX rx6600

Sandisk SATA SSD

Silverstone Sg13 Mini ITX Case

Generic old TT 120mm fan (Soon to be upgraded)

 

Temps for 5600 stock fan curve:

Idle- 42 Degrees

Cinnebench- 72 Degrees 1.05V 3.7Ghz 100% of 76W TDP

Bf2042- 76 Degrees

BF1- 77 Degrees

 

Temps for 3600 Stock fan curve:

Idle- 42-43 Degrees

Cinnebench- 84 Degrees 1.33V 3.9Ghz 91% of 88W TDP

Bf2042- 70 Degrees

 

[Shimejii]

Why are you making a 2nd topic about this? You just made one an hour ago.

[Cantaloupeman]

14 minutes ago, Shimejii said:

Why are you making a 2nd topic about this? You just made one an hour ago.

Oof I didn't think I would get any responses and I figured this was a better place to post it. 

[PDifolco]

3 minutes ago, Cantaloupeman said:

Oof I didn't think I would get any responses and I figured this was a better place to post it. 

Really there isn't much to say... You have pretty comparable perfectly normal temps on both CPU, and it' s normal for the 5600 to get a bit hotter as PBO is more effective

You don't give any figures about perfs except "more stutters", how can we help ?

 

[Mark Kaine]

41 minutes ago, Cantaloupeman said:

91% of 88W TDP

the TDP of the 3600 is 65w.

 

it stands for thermal (or total) design power. its what the chip is "targeting" by design. 

 

yes it can go higher, i think technically the limit is 1000w or something but that is *totally irrelevant* to your issue. 

 

i mean most likely,  because i didn't fully understand what the issue is tbh due to your fixation on "TDP" and stuff. Can you explain the issue in a short sentence perhaps,  what is the actual issue you're seeing that the 5600 performs actually worse?

Did you run any relevant benchmarks,  like superposition or firestrike? 

 

 

also im agreeing with you because my 3600 is also faster than my 5800x3D, in certain applications. So i think this is entirely possible, but i don't think that you have explained the issue well.

 

also kind of feels the issue is probably elsewhere.

 

what about the RAM? XMP? Speed?

Dual channel configuration? 

 

GPU ... temps? (doesn't sound like you have particularly good case cooling?)

 

 

 

15 minutes ago, PDifolco said:

You don't give any figures about perfs except "more stutters", how can we help ?

pretty much this 

[Cantaloupeman]

1 minute ago, Mark Kaine said:

thermal design power. its w

 

3 minutes ago, Mark Kaine said:

the TDP of the 3600 is 65w.

 

it stands for thermal design power. its what the chip is "targeting" by design. 

 

yes it can go higher, i think technically the limit is 1000w or something but that is *totally irrelevant* to your issue. 

 

i mean most likely,  because i didn't fully understand what the issue is tbh due to your fixation on "TDP" and stuff. Can you explain the issue in a short sentence perhaps,  what is the actual issue you're seeing that the 5600 performs actually worse?

Did you run any relevant benchmarks,  like superposition or firestrike? 

 

 

also im agreeing with you because my 3600 is also faster than my 5800x3D, in certain applications. So i think this is entirely possible, but i don't think that you have explained the issue well.

 

also kind of feels the issue is probably elsewhere.

 

what about the RAM? XMP? Speed?

Dual channel configuration? 

 

GPU ... temps? (doesn't sound like you have particularly good case cooling?)

 

 

 

pretty much this 

Okay my bad. 

My issue is that my Temps were higher. Like 10 degrees or more higher in BF2042 when I upgraded to the 5600. Once I put the 3600 back in I noticed my stuttering went away and I gained maybe 10 fps. My clock speeds are lower on the 5600 as well. I'm just concerned there is something wrong with the 5600 I just bought. 

 

My RAM is 3600MHZ DDR4 Kingston. I have XMP enabled. Only one stick of 16gb RAM. GPU temps I haven't paid much attention to because they seemed normal to me. Case cooling isn't the best at the moment. Still doesn't make sense to me how the 5600 can be that much hotter in the same case with the exact same cooler. 

 

The reason I mentioned TDP was because I couldn't get Precision boost to enable in Ryzen Master. I thought maybe the fact that 100% of my TDP was being used on the 5600 meant that it was slowing my clocks down and hurting performance. 

 

I ran superposition but didn't think to record the performance since my CPU temps were only a few degrees hotter than idle. I didn't feel like buying 3dmark for the CPU benchmark either. I just decided to do cinnebench instead. 

[Cantaloupeman]

16 minutes ago, PDifolco said:

Really there isn't much to say... You have pretty comparable perfectly normal temps on both CPU, and it' s normal for the 5600 to get a bit hotter as PBO is more effective

You don't give any figures about perfs except "more stutters", how can we help ?

 

I mean my main concern was temps and that I maybe had a defective processor. Thats why I included the temps I recorded. I couldn't get PBO to enable either so I don't understand why the 5600 was more than 10 degrees hotter than my old CPU and seemed to be a downgrade in game. I got about 10 less frames with the 5600 and it felt extremely stuttery. Idk what else to say

[Mark Kaine]

12 minutes ago, Cantaloupeman said:

 

Okay my bad. 

My issue is that my Temps were higher. Like 10 degrees or more higher in BF2042 when I upgraded to the 5600. Once I put the 3600 back in I noticed my stuttering went away and I gained maybe 10 fps. My clock speeds are lower on the 5600 as well. I'm just concerned there is something wrong with the 5600 I just bought. 

 

My RAM is 3600MHZ DDR4 Kingston. I have XMP enabled. Only one stick of 16gb RAM. GPU temps I haven't paid much attention to because they seemed normal to me. Case cooling isn't the best at the moment. Still doesn't make sense to me how the 5600 can be that much hotter in the same case with the exact same cooler. 

 

The reason I mentioned TDP was because I couldn't get Precision boost to enable in Ryzen Master. I thought maybe the fact that 100% of my TDP was being used on the 5600 meant that it was slowing my clocks down and hurting performance. 

 

I ran superposition but didn't think to record the performance since my CPU temps were only a few degrees hotter than idle. I didn't feel like buying 3dmark for the CPU benchmark either. I just decided to do cinnebench instead. 

ok thats better... i mean it does sound like there's an issue,  so all the relevant info is helpful to figure this out.

 

3D mark has a free demo including firestrike on Steam.  check the store page it should say "demo" right there. 

 

it would be useful because then we can compare the performance to similar systems...

 

8 minutes ago, Cantaloupeman said:

I mean my main concern was temps and that I maybe had a defective processor. Thats why I included the temps I recorded. I couldn't get PBO to enable either so I don't understand why the 5600 was more than 10 degrees hotter than my old CPU and seemed to be a downgrade in game. I got about 10 less frames with the 5600 and it felt extremely stuttery. Idk what else to say

also before you do anything,  uninstall Ryzen Master,  reset BIOS (cause Ryzen Master can actually mess that up) use XMP, but leave everything else at default settings. 

 

Then you could run the firestrike "demo" as Ryzen Master can actually cause issues on some systems so its better to not have that mess with things. its also not really useful outside some niche use cases. 

 

[Cantaloupeman]

6 minutes ago, Mark Kaine said:

ok thats better... i mean it does sound like there's an issue,  so all the relevant info is helpful to figure this out.

 

3D mark has a free demo including firestrike on Steam.  check the store page it should say "demo" right there. 

 

it would be useful because then we can compare the performance to similar systems...

 

also before you do anything,  uninstall Ryzen Master,  rest BIOS (cause Ryzen Master can actually mess that up) use XMP, but leave everything else at default settings. 

 

Then you could run the firestrike "demo" as Ryzen Master can actually cause issues on some systems so its better to not have that mess with things. its also not really useful outside some niche use cases. 

 

Should I put the 5600 back in? What else should I look at before I take the 3600 out?

[Mark Kaine]

4 minutes ago, Cantaloupeman said:

Should I put the 5600 back in? What else should I look at before I take the 3600 out?

... actually you can run firestrike with the 3600... post results here, then put the 5600 back and run it again,  really cant hurt to compare both (results)

other than removing Ryzen Master , resetting BIOS and settings XMP, not really anything else you should do, for now. 

 

Edit: except install  newest AMD chipset drivers... that would probably be good to do right away as well.

 

 

[SansVarnic]

-= Topics Merged =-

Please avoid reposting topics, if you need it moved, report your topic and request the topic be moved.

[Cantaloupeman]

16 minutes ago, Mark Kaine said:

... actually you can run firestrike with the 3600... post results here, then put the 5600 back and run it again,  really cant hurt to compare both (results)

other than removing Ryzen Master , resetting BIOS and settings XMP, not really anything else you should do, for now. 

 

Edit: except install  newest AMD chipset drivers... that would probably be good to do right away as well.

 

 

Isn't the firestrike demo just a gpu benchmark and you have to pay extra for the CPU one?

[Mark Kaine]

42 minutes ago, Cantaloupeman said:

Isn't the firestrike demo just a gpu benchmark and you have to pay extra for the CPU one?

nope , it tests both 

[jaslion]

12 hours ago, Cantaloupeman said:

Cinnebench- 72 Degrees 1.05V 3.7Ghz 100% of 76W TDP

Thats already below base spec

 

It should go to 1.2-1.3v just like the 3600