Overclock.net banner
841 - 860 of 1025 Posts

·
Registered
Joined
·
511 Posts
Produktion Week 2022. My X3D is 2207.
Same batch here. Mine goes down to 2880MHz at stock balanced windows 10 power plan, but most of the time sits at ~3600MHz on idle or low load

 

·
Premium Member
Joined
·
6,914 Posts
If average fps is an .... average of min/max fps in a given time period how could not this improve the average number while increasing the min values? :geek:
Because the min's people use in graphs are just 1% of frames or sometimes 0.1% of frames. It is the lowest drops/stutters in the game, but it makes up very little of the actual frames.

If you have ever been playing a game and most of the time it goes along fine and smooth and then suddenly you notice a single hiccup in the smoothness, that is what those minimums in the graph are.
 

·
Registered
Joined
·
1,506 Posts
I know what a min fps graph is :) I was just poniting out to the "even though ram tuning often doesn't improve avg fps by much on 5800X3D, it can work wonders for min fps " statement that sounded wierd

while this explanation maybe on point "Smaller standard devian from the mean. It spends less time at 99% and less time at 1% so averages remain the same but frame pacing is more consistent providing improved consistency. Pretty much exactly what is observed by most people as smooth game play but doesn't show up well in metrics on charts. " but even here, why it should be reduced its "time (fps level) at 99%" when having a quicker IF and RAM? Unstable overclock?
 

·
Registered
Joined
·
758 Posts
Smaller standard devian from the mean. It spends less time at 99% and less time at 1% so averages remain the same but frame pacing is more consistent providing improved consistency. Pretty much exactly what is observed by most people as smooth game play but doesn't show up well in metrics on charts.

EDIT: I do appreciate everyones input in online reviews, I'm not maliciously attacking jufus but...

I wish jufus would post SotTR 1080 lowest with SMT off as a measuring stick for comparison to determine the efficacy of his OCing. I doubt his x3d setup would crack 370fps.

It's another data point but, IMHO Jufus is, at best, mediocre when tuning Zen chips, advocating for static OCs as the best method, so his AMD CPU results should be viewed as the equivalent of low effort CPU OCing.
View attachment 2560237
In his x3d vs AL video he was allegedly running at 4.7ghz on x3d but covered all info except latency on his Aida ram screenshot. 57.4 ns suggest loose timings for a 4.7 OC or massive clock stretching. 3800C14 for me, returns 56.2ns at 4610 all core speeds. 4112 C15 is 53.2ns. I mention this as, if this is his max tuned setup, it creates doubt as to his actual CPU OC efficacy. Notice he showed zero productivity tests as a result of his OC. R23 should have been 16k. You think he scores 16k....???

He claims cooling issues, but during gaming I never see my CPU over 42c CCD1. Add 20c to eliminate chiller and...62c, not exactly hot. Suggests he's brute forcing stability with too much voltage. I know from testing that is the wrong way to OC x3d and leads to stock level performance as FIT throws the brakes on hard.

Compared to a self proclaimed golden sample 12900k running max tuned DDR 5 7000+ with 32 CL and 4133 C15 on a different 12900k that has great DDR4 IMC.
I agree his tuning is avg at best, but his point on ramtuning making some games run smoother with much higher lows after ram tuning stands. I don't think he's right about pbo+co being inferior to static clocks.
 

·
Registered
Joined
·
511 Posts
  • Rep+
Reactions: ice445

·
Registered
Joined
·
865 Posts
Discussion Starter · #848 ·
I know what a min fps graph is :) I was just poniting out to the "even though ram tuning often doesn't improve avg fps by much on 5800X3D, it can work wonders for min fps " statement that sounded wierd

while this explanation maybe on point "Smaller standard devian from the mean. It spends less time at 99% and less time at 1% so averages remain the same but frame pacing is more consistent providing improved consistency. Pretty much exactly what is observed by most people as smooth game play but doesn't show up well in metrics on charts. " but even here, why it should be reduced its "time (fps level) at 99%" when having a quicker IF and RAM? Unstable overclock?
I suspect that's exactly why, yes. With my own experimentation, I've found X3D is highly sensitive to voltage, too much or too little, and performance tanks, but not universally across benchmarks. Different from Zen 3 X, mostly because of the FIT cap of 1.3 resulting in a drastically narrowed window of "ideal" according to FIT.

Don't gete wrong, Jufus does some testing other channels simply don't, so I value him as another information resource. The reality is he simply doesn't invest the time to test a huge number of configurations and find which provides the best balanced performance. I don't even fault him for that, he readily admits to breng a gamer first Overclocker second.

I've learned my early testing, all the Hwbot scores from April had garbage efficiency due to clock stretching, however beating them still isn't easy. 162xx CB23 16 thread for example and 1597 single thread. I have a number of bios profiles saved that should match that, yet they do not due to voltage mismatch issues Ive not yet 100% resolved. It takes way more finesse to OC X3D with max efficiency than X CPUs imho when surpassing 4650 all core boost
 

·
Registered
Joined
·
4 Posts
I still think something seems off with my 5800x3d.
Maybe I am just paranoid something is messed up. Even while gaming or benching the cores will not stay at max frequency.

How about everyone elxe? Do your clockspeeds bounce up and down like this under heavy load with thermals in check. While gaming and benching it does not go over 75c. Here is a picture of MSI afterburner. The bios is at default accept the Ram OC. I am using the beta bios for 1207. It acted the same in the old bios 1206b

As you can see, it bounces between 3.6 GHz and 4.45 /4.5.
Using max 85w

Rectangle Font Slope Parallel Symmetry
 

Attachments

·
Registered
Joined
·
758 Posts
I suspect that's exactly why, yes. With my own experimentation, I've found X3D is highly sensitive to voltage, too much or too little, and performance tanks, but not universally across benchmarks. Different from Zen 3 X, mostly because of the FIT cap of 1.3 resulting in a drastically narrowed window of "ideal" according to FIT.

Don't gete wrong, Jufus does some testing other channels simply don't, so I value him as another information resource. The reality is he simply doesn't invest the time to test a huge number of configurations and find which provides the best balanced performance. I don't even fault him for that, he readily admits to breng a gamer first Overclocker second.

I've learned my early testing, all the Hwbot scores from April had garbage efficiency due to clock stretching, however beating them still isn't easy. 162xx CB23 16 thread for example and 1597 single thread. I have a number of bios profiles saved that should match that, yet they do not due to voltage mismatch issues Ive not yet 100% resolved. It takes way more finesse to OC X3D with max efficiency than X CPUs imho when surpassing 4650 all core boost
I wait for i2hard to do their 5800X3D-test, they usually tweak close to the edge and tune Zen better than Jufes.
 

·
Registered
Joined
·
488 Posts
I still think something seems off with my 5800x3d.
Maybe I am just paranoid something is messed up. Even while gaming or benching the cores will not stay at max frequency.

How about everyone elxe? Do your clockspeeds bounce up and down like this under heavy load with thermals in check. While gaming and benching it does not go over 75c. Here is a picture of MSI afterburner. The bios is at default accept the Ram OC. I am using the beta bios for 1207. It acted the same in the old bios 1206b

As you can see, it bounces between 3.6 GHz and 4.45 /4.5.
Using max 85w

View attachment 2560300
Yeah, that's not right. I didn't see any reductions in clock speed when I ran Cinebench R23, and my reported PPT peaked around 123W. Maybe a motherboard setting that needs tweaking? A lot of motherboards have some sort of "performance booster" that's usually active by default that in my experience messes things up.
 

·
Registered
Joined
·
13 Posts
I was bothered that my board wouldn't hit that 45.5x Multiplier so I flashed the 16B BIOS back on the board. It now boosts as expected. Out of curiosity I wanted to see if I could get 2000 1:1:1 stable and...

It runs Prime 95... It runs BOINC...

It just spits out WHEA errors at regular ticks. VSOC 1100mv - VDDP 920mv, VDDG CCD 950mv, VDDG IOD 980mv

When I tried to change VDDG IOD to 1050mv it spit out WHEA errors faster. When I tried to drop the voltage to 950mv it booted into a BSOD...

PLL is already at 1.964

So it's stable but not stable?

Edit: Ram is just at XMP to test for FLCK stability.

Rectangle Font Screenshot Parallel Technology

Sky Font Symmetry Magenta Darkness

Font Screenshot Display device Multimedia Electronic device
 

·
Registered
Joined
·
2,978 Posts
@Sparrow1408 You need to look at Event Viewer and create a custom view for WHEA errors

Update on my chip. I got another WHEA error last night 05/11. That's 10 days after my first one on May 1st, and that's ~20 days after installing the chip. Seems like it throws a WHEA error every 10 days. That's only 2 since I installed the chip and have been running it at 1900:3800. Tested 1933, threw a bunch of errors so it seems 1900 is at the edge just before stability.

EDIT: I guess I got my 1900:3800 Club card revoked

Odd thing is I can't reproduce any WHEA errors while running OCCT, y-cruncher, TM5
 

·
Registered
Joined
·
38 Posts
Try 1,13 VSOC
 

·
Registered
Joined
·
2,978 Posts
Turns out the reason I couldn't boot at 1866:3733 with manual PCB rev timing suggestions from Ryzen Calculator was the tCWL was a mis-match with CL. tCWL was suggested to be 12 while CL could only boot at 14. From what I gathered tCWL has to be = to or > CL If this is incorrect let me know. Sorry for off-topic
 

·
Registered
Joined
·
286 Posts
tcwl is also dependend on trdwr its kinda a tradeoff but mostly its worth rather to run it 1:1 with CL and reduce trdwr as much as possible.
 

·
Registered
Joined
·
2,978 Posts
Does HWinfo64 catch all the WHEAs?
Good question. I always assumed it did but I always relied on Event Viewer as it's a windows tool. The errors I caught were not when I was running OCCT, y-cruncher, nor TM5. Oddly they were 10 days apart. Each of them. And Event Viewer will always catches them no matter what. I would assume HWinfo would need to be running to catch them.
 

·
Registered
Joined
·
51 Posts
Good question. I always assumed it did but I always relied on Event Viewer as it's a windows tool. The errors I caught were not when I was running OCCT, y-cruncher, nor TM5. Oddly they were 10 days apart. Each of them. And Event Viewer will always catches them no matter what. I would assume HWinfo would need to be running to catch them.
I run HWinfo64 all the time and haven't noticed any yet at 1866 fclk.
 
841 - 860 of 1025 Posts
Top