AMD Ryzen 7 5700X

[Mussels]

$369 Au for them here and i think they're great at that price, since they're capable of running any GPU currently out there happily

 

[amd64skater]

I've had the Ryzen 7 5700x for about 5 months now and I am incredibly happy with it's performance. Just like its big brother the 5800x it's toe to toe basically. Yeah it falls short but barely in some benchmarks to the 5800x. This this processor is amazing and I'm glad I went this route.

 

[Mussels]

I've had the Ryzen 7 5700x for about 5 months now and I am incredibly happy with it's performance. Just like its big brother the 5800x it's toe to toe basically. Yeah it falls short but barely in some benchmarks to the 5800x. This this processor is amazing and I'm glad I went this route.

It's a 5800x with Eco mode, or more accurately the 5800x is the 5700x with the power limits removed

 

[Mase]

From the CPU Temperatures shown below, would using a 5800X in eco mode (via PBO) achieve the same or similar temperature as the 5700X at stock (45C)?

I'm guessing the 5700X at stock has PBO disabled, so the power draw would be capped.

Whereas the 5800X would have PBO enabled in eco mode, so I'm guessing this means it would boost above the 'power limit'.

I only ask as I would like to achieve the 5700X 45C temps shown below, but the 5800X is currently available at £220 vs the 5700x costing £250.

So I'm trying to figure out if I can easily achieve the stock 5700X temps with the 5800X without having to manually fiddle with voltages and test stability etc.

 

[Colddecked]

From the CPU Temperatures shown below, would using a 5800X in eco mode (via PBO) achieve the same or similar temperature as the 5700X at stock (45C)?

I'm guessing the 5700X at stock has PBO disabled, so the power draw would be capped.

Whereas the 5800X would have PBO enabled in eco mode, so I'm guessing this means it would boost above the 'power limit'.

I only ask as I would like to achieve the 5700X 45C temps shown below, but the 5800X is currently available at £220 vs the 5700x costing £250.

So I'm trying to figure out if I can easily achieve the stock 5700X temps with the 5800X without having to manually fiddle with voltages and test stability etc.

View attachment 267794

Yes if you use it in eco mode it will behave exactly like a 5700x.

 

[Mase]

Thanks - I'm not certain but I think a 5800X in eco mode might be equivalent to a 5700X with PBO enabled.

So I'm guessing the temps would be a bit higher than the 5700X at stock (45C) shown in the table above - assuming the 5700X at stock has PBO set to auto/disabled.

 

[Colddecked]

Thanks - I'm not certain but I think a 5800X in eco mode might be equivalent to a 5700X with PBO enabled.

So I'm guessing the temps would be a bit higher than the 5700X at stock (45C) shown in the table above - assuming the 5700X at stock has PBO set to auto/disabled.

No if you set the 5800x to the same power settings as the 5700x it will act exactly like a 5700x, maybe slightly higher clocks since the bin is supposed to be better.
Just buy the 5800x it costs less its a no brainer.

 

[Mussels]

From the CPU Temperatures shown below, would using a 5800X in eco mode (via PBO) achieve the same or similar temperature as the 5700X at stock (45C)?

I'm guessing the 5700X at stock has PBO disabled, so the power draw would be capped.

Whereas the 5800X would have PBO enabled in eco mode, so I'm guessing this means it would boost above the 'power limit'.

I only ask as I would like to achieve the 5700X 45C temps shown below, but the 5800X is currently available at £220 vs the 5700x costing £250.

So I'm trying to figure out if I can easily achieve the stock 5700X temps with the 5800X without having to manually fiddle with voltages and test stability etc.

View attachment 267794

Not quite, the 5800x still runs hotter. I've compared both and the 5700x runs colder because it doesnt boost as high.


It would be possible to tune a 5800x with custom PBO limits, a negative offset and so on but it would take more effort than just eco on.

No if you set the 5800x to the same power settings as the 5700x it will act exactly like a 5700x, maybe slightly higher clocks since the bin is supposed to be better.
Just buy the 5800x it costs less its a no brainer.

you cant set the 5800x and 5700x to use the same clock speeds, so they behave differently. a -200 offset on the 5800x might get them close, but i have no 5700x to test that comparison these days.

 

[amd64skater]

When I bought my 5700x it was $240. The 5800x was $389. For 100mhz slower performance and way cooler temps. I'm still happy to be a proud 5700x owner.

 

[Mussels]

When I bought my 5700x it was $240. The 5800x was $389. For 100mhz slower performance and way cooler temps. I'm still happy to be a proud 5700x owner.

I dont own one and i'm STILL proud of the 5700x. It should have been launched on day one, instead of this trend of maximum wattage for tiny gains.


the 5800x3D feels like a 5700x with the 3D cache on top, with the lower wattages, temps and clock speeds vs the 5800x (140W to 120W at peak, but gaming wattages and temps are right down there with the 57x)


5700x aint on this graph, but you can easily imagine where it slots in

 

[Mase]

Great replies, thanks.

I'm leaning towards the 5700X even though it is £30 more expensive as odd as that sounds.

The performance difference to the 5800X is negligible to me as I really won't notice any difference, the temps however look like a stark contrast.

Sounds like I could tweak the 5800X to behave similarly but like you say this would take effort as opposed to just plug and play.

I wonder if these settings for the 5800X (PPT = 61W, TDC = 45A, EDC = 65A) would put it in the same temps ball park as the 5700X at stock.

 

[Mussels]

Great replies, thanks.

I'm leaning towards the 5700X even though it is £30 more expensive as odd as that sounds.

The performance difference to the 5800X is negligible to me as I really won't notice any difference, the temps however look like a stark contrast.

Sounds like I could tweak the 5800X to behave similarly but like you say this would take effort as opposed to just plug and play.

I wonder if these settings for the 5800X (PPT = 61W, TDC = 45A, EDC = 65A) would put it in the same temps ball park as the 5700X at stock.

If the 5800x is cheaper, absolutely snag it and work on undervolting it. I made an entire thread about recommended settings, but the one thing i never tried was negative clockspeed offsets since i was chasing those sweet GHz

The settings i used were:
PPT: 95W
TDC:60A
EDC: 90A

They had me under 60C at full load, vs the normal 80C

throw -100 or -200 on top of that with curve optimiser and you'd beat a 5700x for stock efficiency

 

[Mase]

If the 5800x is cheaper, absolutely snag it and work on undervolting it. I made an entire thread about recommended settings, but the one thing i never tried was negative clockspeed offsets since i was chasing those sweet GHz

The settings i used were:
PPT: 95W
TDC:60A
EDC: 90A

They had me under 60C at full load, vs the normal 80C

throw -100 or -200 on top of that with curve optimiser and you'd beat a 5700x for stock efficiency

Ok I'll give it a go, is there a chance I would run into stability issues using those settings?

 

[Mussels]

Ok I'll give it a go, is there a chance I would run into stability issues using those settings?

zero

limiting the power doesnt alter voltages, so it cannot harm stability, only performance


Using the curve undervolt can go unstable, but since it uses less voltage/wattage, you get more performance within those limits - it's the second step, not the first

 

[Kodehawa]

There is no need to burden yourself with proving that, as there is no B0 stepping 5700X, this is a newly released SKU. The closest thing to it would be a B0 5800X in eco mode, which is essentially what this processor is.

My 5950X is a B0, I bought it a few months before the B2 stepping was announced by AMD. It's about as tightly binned as a Ryzen can be

My 5700X is B0! The date on the CPU itself is actually 2043 (43th week of 2020).

Validator link: https://valid.x86.fr/cbnkse

Just felt like making (actually, logging into a really old one) an account to say, well, B0 5700X exists, even at retail. I find it weird, as I, myself, don't find any B0 5700X references online and all pictures I find are B2.

I have the box and all, and everything matches up/passes AMD verification as a 5700X. Even most pictures I see of 5700X online have something like 2212 as the date, which makes a lot more sense than 2043, but I can assure you it is 2043, in a kind-of blurry image I took when I got it.

http://imgur.com/Nw2EHhI

 

[Mussels]

Bit blurry but thanks for posting that

Any obvious differences? Dual CCX design with one disabled like some other CPUs had?

 

[The King]

Bit blurry but thanks for posting that

Any obvious differences? Dual CCX design with one disabled like some other CPUs had?

Yes, my thought as well possible early 5700X that was made from a defective higher B0 SKU.

 

[Kodehawa]

Yes, my thought as well possible early 5700X that was made from a defective higher B0 SKU.

Doesn't seem to be dual CCD. First CCD is CCD1 on HWInfo and CTR doesn't say anything about it either (as in the "special processor" string), I think.
Attaching CTR logs just in case. I don't really know what to look for here

***ClockTuner for Ryzen 2.1 RC5(ver.23) by 1usmus***
AMD Ryzen 7 5700X 8-Core Processor
ASUS TUF B450M-PLUS GAMING
BIOS ver. 3802 SMU ver. 56.70.00
TABLE ver. 3672325
DRAM speed 3400 MHz
12/07/2022 10:38:07


P1
FREQ CCX1# 0MHz
VID# 1250mV
Usage trigger# 81%
OB CCX1# 1
OB CCX2# 1
P2
FREQ CCX1# 0MHz
VID# 1350mV
Usage trigger# 56%
OB CCX1# 1
OB CCX2# 1
PX
FREQ HIGH# 0MHz
FREQ MID# 0MHz
FREQ LOW# 0MHz
VID HIGH# 1375mV
VID MID# 1375mV
VID LOW# 1350mV
OB HIGH# 1
OB MID# 1
OB LOW# 1
Usage trigger# 18%
PX OB LIMIT# 75MHz

Also attaching CPU-Z log.

 

[The King]

Doesn't seem to be dual CCD. First CCD is CCD1 on HWInfo and CTR doesn't say anything about it either (as in the "special processor" string), I think.
Attaching CTR logs just in case. I don't really know what to look for here



Also attaching CPU-Z log.

Use Zentimings and use the debug report option.
Although I think CTR would have detected if there were two CCDs but not sure has I don't use it.

 

[Kodehawa]

Use Zentimings and use the debug report option.
View attachment 273395

It reports:

CCDCount: 1
CCXCount: 1
NumCoresInCCX: 8

Attaching:

 

[The King]

It reports:


Attaching:

DDR4 3400MT is an odd frequency to be running on ZEN3? Why not 3600 or 3800? Post your Zentimings

OK, I see it was in the report your XMP is 3200 CL16-20-20-20? and you running 3400 CL16-20-20-20.
May have a problem going to 3600 without increasing to something like 3600 CL16-21-21-21.

 

[Kodehawa]

DDR4 3400MT is an odd frequency to be running on ZEN3? Why not 3600 or 3800? Post your Zentimings

OK, I see it was in the report your XMP is 3200 CL16-20-20-20? and you running 3400 CL16-20-20-20.
May have a problem going to 3600 without increasing to something like 3600 CL16-21-21-21.

I remember trying 3600 on my Zen+ and it not being stable at default XMP timings (failed most memory tests 4~6h into them). Haven't tried on Zen 3 yet, it might work or no, I'll give it a shot later after exams end.

I pushed it to 3400 here (again) so the IF was 1:1, it wasn't at 3200.

ZenTimings:

http://imgur.com/0eu7e4s

 

[Veseleil]

Since I'm watching this thread from the start, but haven't noticed actual 5700X owners comments on max stable FCLK frequency, my question will be obvious: How high FCLK your 5700X can achieve?
I want to upgrade my 3600, and i need to know beforehand. If it can't hit stable at least 1900, it's a real bummer, and I'll gladly consider other options then.

 

[The King]

Since I'm watching this thread from the start, but haven't noticed actual 5700X owners comments on max stable FCLK frequency, my question will be obvious: How high FCLK your 5700X can achieve?
I want to upgrade my 3600, and i need to know beforehand. If it can't hit stable at least 1900, it's a real bummer, and I'll gladly consider other options then.

If the CPU does not have the dreaded FCLK 1900 hole depending on what motherboard and memory kit you buy you can easily go over 2000 FCLK with almost any ZEN 3 CPU with two
8GB single rank DIMMS.

Increasing to Dual rank DIMMS will usually make higer FCLK more difficult to achieve but not impossible if you know what settings to change.

I remember trying 3600 on my Zen+ and it not being stable at default XMP timings (failed most memory tests 4~6h into them). Haven't tried on Zen 3 yet, it might work or no, I'll give it a shot later after exams end.

I pushed it to 3400 here (again) so the IF was 1:1, it wasn't at 3200.

ZenTimings:

http://imgur.com/0eu7e4s

You may need to increase voltage as well if you go to 3600MT. you can try with the settings I gave 3600 CL16-21-21-21 even try 16-22-22-22 and if your XMP is 1.35V then try 1.4V VDIMM.
If not stable then just run it the way you have it now. If you are not comfortable or familiar with these settings best leave it has it is.

Rather not break you PC before exams and blame me.

 

[Kodehawa]

If the CPU does not have the dreaded FCLK 1900 hole depending on what motherboard and memory kit you buy you can easily go over 2000 FCLK with almost any ZEN 3 CPU with two
8GB single rank DIMMS.

Increasing to Dual rank DIMMS will usually make higer FCLK more difficult to achieve but not impossible if you know what settings to change.


You may need to increase voltage as well if you go to 3600MT. you can try with the settings I gave 3600 CL16-21-21-21 even try 16-22-22-22 and if your XMP is 1.35V then try 1.4V VDIMM.
If not stable then just run it the way you have it now. If you are not comfortable or familiar with these settings best leave it has it is.

Rather not break you PC before exams and blame me.

Is there a significant performance improvement by doing so? I'll give it a try after the 19th if so