Ryzen 7 2700 seriously slow

[Indus]

I got the CPU new on sale for $200 vs $320 for the X. I wanted a full feature MOBO and one with reliable circuitry that could operate for many years. Research suggested most boards came with questionable or poor VRMs and heatsinks that could be problematic long term (especially in my situation where it's in a hot environment). In my budget it seemed only some MSI and the ASRock board had a proper VRM to handle the 8 cores, and the MSI boards were supposed to be the ones with BIOS issues (like high voltage [1.5v+] w/ limited adjustment options).

So the choice was $460 for an X CPU and middling MOBO or $400 for the non X and a supposedly great MOBO (everyone claimed it had voltage, XFR/PBO adjust etc). Based on my research it sounded like an easy thing to get a 2700 to operate like a 2700x, with proper boost etc. The savings would also allow me to get a better heatsink (even the crappy 212 Evo is notably better than the good stock cooler of the 2700x) plus a full set of good new case fans (apparently replacing the 212 Evo fan with these made it much better). Certainly at regular retail pricing I would've spent the $30 more for the 2700x (and probably tried it's stock cooler).

And I did get 3200MHz RAM...2400MHz is just what the MOBO chose to run at "default" settings. Got it for a great deal too (especially compared to even just 6-12 months ago). If you read the thread that information is clearly posted (my first post even).

If you can, just sell the 2700 and get a 2600x. You'd probably make money on that deal and just watch the defaults give you higher clocks. Or you can get a 2700x.

Regarding memory clocks.. you do have to run the XMP setting in bios.

 

[LTC8K6]

I was going off of what I read in reviews such as here:



If that's wrong, then fine. Either way we're back to his CPU shouldn't be locked to 3.2 GHz and either his board is faulty or Asrock completely screwed up the BIOS for non-X CPUs.

Here's a review where the Asrock Taichi U seems to perform just as well as an ASUS board with the 2700 when not overclocked:

https://wccftech.com/review/amd-ryzen-7-2700-cpu-review-asrock-x470-asus-x470-motherboards/9/

They don't specify what clocks they're seeing from what I saw skimming through the review, but I doubt the 2700 on the Asrock board would match the ASUS board or beat the 1700x they have if the Asrock board limited the 2700 to 3.2 GHz.

"in the form" ...the XFR2 doesn't boost as high with the 2700, that's all. It's there though, as is PB2.
What's going on with the OP's system looks like it will be an unanswered question apart from a general "the boost is not working".

 

[TheELF]

Gamers nexus just made a video pretty much confirming everything he said,it might be rare or only with certain mobos but it certainly happens.
In the video the 2700 is stuck at 3.4 mostly but it's with a game so a lighter load which makes the OPs claim of 3.2 in a benching app very credible.

 

[IEC]

3.2GHz is the base clock without PBO enabled. Stock 2700 runs 3.4GHz boost with a little additional from XFR if your cooling/power limit allows. That's why GN got 3.467GHz on all cores.

In this chart, we ran the test on the R5 2600 versus the R7 2700, both stock, and logged the frequency on each core during the test. As illustrated by the chart, the R5 2600 maintains a much higher frequency of 3816 to 3850MHz all-core, something we can see as plotted against all 6 core clocks in this chart. The R7 2700 runs closer to 3467MHz all-core. Although we like recommending the R7 2700 non-X, it’s normally under the stipulation of a user-applied overclock. There’s still a lot of all-core headroom to gain here, but this illustrates exactly why the R7 2700 gets outmatched by the cheaper R5 2600 in frequency-bound titles.

Source: https://www.gamersnexus.net/hwreviews/3474-new-cpu-testing-methodology-2019-ryzen-3000-prep

 

[Topweasel]

I am sure I have read people suggesting it before. But on my 1700 3GHz is the base clock basically what no matter what a core will run at with all self or manually set turbos and over-clocks not set. The user is not seeing a CPU temp this is near any limit and to a degree relatively cool. Since 3.2 is the base for 2700. The only thing I can think of is that manual overclock is enabled but that all of the settings are left at stock. This will disable all the XFR and PBO functionality while not having it clocked higher because again default settings for the manual overclock.

Has the OP actually tried to just plain reset the BIOS to stock.

 

[VirtualLarry]

Has the OP actually tried to just plain reset the BIOS to stock.

I assume that you mean, using the actual physical "CLR_CMOS" jumper on the board?

 

[EliteRetard]

Yeah the 2700 is still a seriously slow, stupid CPU. I don't see how anyone can argue that. AMD's second best CPU is slower than their low tier mid range CPU. Requiring people to overclock to get the advertised speed/performance is just broken. It should be 4.1GHz boost, not 3.45GHz under lighter loads...and under heavy loads it shouldn't be 3.2GHz, it should be like 3.8GHz...200-300MHz less than the 2700x as advertised. And MOBO makers should not be removing PBO/XFR options and preventing them from adjusting power limits and boost clocks.

With all the hype around Ryzen, tons of recommendations for the 2700, all the rage about how great PBO/XFR boost the chips...it was not clear anywhere I researched at the time how seriously slow and broken the 2700 was. Or how broken the ASRock board is for non X CPUs. Just a huge and frustrating disappointment.

Somebody tried to compare it to Intel earlier, saying that nobody would complain if Intel did it or some nonsense...(obviously I must be an Intel shill) Imagine if the 9900 non K only ran at 3.45GHz under light loads (performing worse than even a 9600 in such scenarios). While a 9600 non K actually boosted properly near 4.5GHz under light loads and held near 4GHz full core stress testing. Anybody who then claimed the 9900 non K was terrible and slow would be entirely correct and fully justified for saying so. Trying to blame them for it would be just as crap.

 

[VirtualLarry]

Requiring people to overclock to get the advertised speed/performance is just broken. It should be 4.1GHz boost, not 3.45GHz under lighter loads...and under heavy loads it shouldn't be 3.2GHz, it should be like 3.8GHz...200-300MHz less than the 2700x as advertised. And MOBO makers should not be removing PBO/XFR options and preventing them from adjusting power limits and boost clocks.

Basically, none of this is true.

I proved that my 2700 CPU does boost to 4.1Ghz under lighter loads, and I don't recall ANYWHERE the 2700 being advertised as "200-300Mhz slower than 2700X in all-core loads"., It IS advertised as a 65W TDP CPU. That much is true. Your imaginative claims as to how it's advertised, I don't know what to say. Other than, maybe provide screenshots of these claims?

 

[Elfear]

Yeah the 2700 is still a seriously slow, stupid CPU. I don't see how anyone can argue that. AMD's second best CPU is slower than their low tier mid range CPU. Requiring people to overclock to get the advertised speed/performance is just broken. It should be 4.1GHz boost, not 3.45GHz under lighter loads...and under heavy loads it shouldn't be 3.2GHz, it should be like 3.8GHz...200-300MHz less than the 2700x as advertised. And MOBO makers should not be removing PBO/XFR options and preventing them from adjusting power limits and boost clocks.

Did you really think a 65W CPU would boost within 200-300Mhz of a 105W CPU? That's like a 5-7% reduction in speed for a massive 38% reduction in power.

The 2700 is very similar to the 1st gen 1700. It performed okay at stock speeds but I think most buyers got it on the cheap in anticipation of overclocking it (at least that's what I did).

 

[Campy]

Overclock it. Not like it's gonna pull 150W at idle just because you overclocked it.

 

[UsandThem]

TLDR:
Spent $1K on new system, got noticeable downgrade, I AM DISSAPOINT!

Oh and I know the Ryzen 3000 (Zen2) are coming out soon, but old system couldn't hold out any longer.

From how this quoted portion is worded, it sounds like you built this PC not that long ago.

I'm just curious if you read any reviews of the CPU before you bought it? Every review I read (most were in the May - June 2018 time-frame) show the performance deficit compared to the 2700X (it is a 65w CPU after all). If you did read reviews before buying it, why are you so disappointed? To me it's a case of caveat emptor. If you wanted the stock performance of the 2700X, you should have spent the extra $50 to get it instead of come here a create this mess of a thread.

https://www.tomshardware.com/reviews/ryzen-7-2700-2700x-review,5606-9.html


https://www.techpowerup.com/reviews/AMD/Ryzen_7_2700/20.html

 

[AnnoyedGrunt]

Seems like there are a couple things going on here.

First, regarding windows boot times, application loading, etc: none of those issues are related to the clock speed. You should be able to run these chips at 2.0 GHz and with any SSD have near instant loading of web pages and such. Those issues to me sound like a software issue.

I recently built a 2700x system with an ASRock B450 mini-itx motherboard and my biggest issue was getting drivers loaded in windows. My first time through I was manually installing video drivers when Windows decided to automatically install an older version. My manual install aborted part way through and the system was laggy and would crash very often.

I reinstalled windows with the network disconnected and that fixed everything. Is it possible that something like that happened?

Regarding power use, I run windows in high performance mode with the CPU minimum set to 5%. Most of the time the CPU is running at 2.1 GHz or so but will ramp up to 4 GHz under load.

As far as the 2700 non-X goes, that chip is somewhat weak out of the box due to the 65 watt TDP. If you overclock it however, you can get roughly the same performance as a 2700x (or more depending on how far you push it). So, you really need to overclock it for the most benefit. I don’t think it should hurt power consumption too much when it is not under load.

-AG

 

[B-Riz]

OP, did you decide what you want to do?

Try a new board? Or just sell everything and see how Zen2 / Intel CL+++++++ turns out?

 

[alexruiz]

OP.
Get a 2700x instead, sell the 2700.
end of thread

 

[beginner99]

First, regarding windows boot times, application loading, etc: none of those issues are related to the clock speed. You should be able to run these chips at 2.0 GHz and with any SSD have near instant loading of web pages and such. Those issues to me sound like a software issue.

Agree. The chips would have to be <1 ghz or somethugn for that to happen which makes me beleive the actual culprit is the ssd. Didn't read whole thread, but that's what I would do: Clone ssd to another ssd and then recheck the system.

 

[TheELF]

Agree. The chips would have to be <1 ghz or somethugn for that to happen which makes me beleive the actual culprit is the ssd. Didn't read whole thread, but that's what I would do: Clone ssd to another ssd and then recheck the system.

No matter what your clocks are at right now,get into bios and turn them down by just 30% and tell us that you didn't notice your system being slower...

 

[Topweasel]

I assume that you mean, using the actual physical "CLR_CMOS" jumper on the board?

Or holding insert at boot up. But yes reset bios not by hand (re not to what the OP thinks it was set to), but to factory defaults.

 

[biodoc]

And MOBO makers should not be removing PBO/XFR options and preventing them from adjusting power limits and boost clocks.

In the bios, go to advanced settings, then AMD CBS, then ZEN common options.

EDIT: I missed a previous post where these options are disable in "non-X" processors.

 

[killster1]

Yeah the 2700 is still a seriously slow, stupid CPU. I don't see how anyone can argue that. AMD's second best CPU is slower than their low tier mid range CPU. Requiring people to overclock to get the advertised speed/performance is just broken. It should be 4.1GHz boost, not 3.45GHz under lighter loads...and under heavy loads it shouldn't be 3.2GHz, it should be like 3.8GHz...200-300MHz less than the 2700x as advertised. And MOBO makers should not be removing PBO/XFR options and preventing them from adjusting power limits and boost clocks.

With all the hype around Ryzen, tons of recommendations for the 2700, all the rage about how great PBO/XFR boost the chips...it was not clear anywhere I researched at the time how seriously slow and broken the 2700 was. Or how broken the ASRock board is for non X CPUs. Just a huge and frustrating disappointment.

Somebody tried to compare it to Intel earlier, saying that nobody would complain if Intel did it or some nonsense...(obviously I must be an Intel shill) Imagine if the 9900 non K only ran at 3.45GHz under light loads (performing worse than even a 9600 in such scenarios). While a 9600 non K actually boosted properly near 4.5GHz under light loads and held near 4GHz full core stress testing. Anybody who then claimed the 9900 non K was terrible and slow would be entirely correct and fully justified for saying so. Trying to blame them for it would be just as crap.

i think you did your research wrong. there is no such drive as 480GB HP EX920 right? so maybe some other things are incorrect too? the 2700 is not the second best cpu from amd right? not even the 10th best i imagine. Feel like you have another issue with your board or psu that is holding you back. get a different psu properly hooked up then switch to a new board.. you can just wait for zen3700x or what ever too but id fix this one first.

 

[Wuzup101]

Yeah the 2700 is still a seriously slow, stupid CPU. I don't see how anyone can argue that. AMD's second best CPU is slower than their low tier mid range CPU. Requiring people to overclock to get the advertised speed/performance is just broken. It should be 4.1GHz boost, not 3.45GHz under lighter loads...and under heavy loads it shouldn't be 3.2GHz, it should be like 3.8GHz...200-300MHz less than the 2700x as advertised. And MOBO makers should not be removing PBO/XFR options and preventing them from adjusting power limits and boost clocks.

With all the hype around Ryzen, tons of recommendations for the 2700, all the rage about how great PBO/XFR boost the chips...it was not clear anywhere I researched at the time how seriously slow and broken the 2700 was. Or how broken the ASRock board is for non X CPUs. Just a huge and frustrating disappointment.

Somebody tried to compare it to Intel earlier, saying that nobody would complain if Intel did it or some nonsense...(obviously I must be an Intel shill) Imagine if the 9900 non K only ran at 3.45GHz under light loads (performing worse than even a 9600 in such scenarios). While a 9600 non K actually boosted properly near 4.5GHz under light loads and held near 4GHz full core stress testing. Anybody who then claimed the 9900 non K was terrible and slow would be entirely correct and fully justified for saying so. Trying to blame them for it would be just as crap.

I don't know what to tell you man. I have a 2700 + X470 Taichi just like you. I upgraded from a 2500K system, just like you (although mine was clocked a bit higher). I'm currently sitting at 4ghz stable on all cores, at 1.35V in Bios and LLC5 (lowest level). Chip is in the low 70s running 16 threads of Prime 95 small FFT max heat. This system feels a LOT faster that my 2500k did. I assume some of that is going from a 1TB SATA SSD to a 1TB NVME SSD. Maybe try overclocking the system using a fixed speed? It's not like it's consuming a ton of power at idle. HW monitor shows values <10w. Package draw running prime is around 110-115W

I haven't tried using PBO/XFR or any of the auto overclocking features of the board yet. Maybe that'll come later, but I figured I'd start out with an old school OC to gather some information about the chip frequencies / voltages first.

 

[maddogmcgee]

Well if nothing else positive came from this thread at least I know I should probably get the 3800x rather than the 3700x (will wait for reviews to confirm).

 

[VirtualLarry]

I did something similar with my 2700, ASUS ROG STRIX B450-F ATX, 16GB GSkill RGB DDR4-3600. I have the CPU at 4.0GHz, 1.3225 or something like that, I don't think I have it as high as 1.350V, will have to check. RAM is at 3400 right now. It has been crashing, however, about once a week, hard freeze more or less. Unsure why. Not mining, have a GTX 1050 3GB connected, not using it for anything other than display.

 

[jaju123]

I did something similar with my 2700, ASUS ROG STRIX B450-F ATX, 16GB GSkill RGB DDR4-3600. I have the CPU at 4.0GHz, 1.3225 or something like that, I don't think I have it as high as 1.350V, will have to check. RAM is at 3400 right now. It has been crashing, however, about once a week, hard freeze more or less. Unsure why. Not mining, have a GTX 1050 3GB connected, not using it for anything other than display.

Have you robustly tested your ram stability? 3400 is very hard to make fully stable. Even my 3600 cl15 b-die kit doesn’t like going above 3333mhz with any timings really.

 

[VirtualLarry]

Well, the timings are really loose, like 18-22-22-40 or something. But you're right, it could be the RAM. I tested it for a week at the default XMP settings that the ASUS UEFI set (3200Mhz), and it seemed stable, then I bumped it at the last moment to 3400, before I put the PC where it is now. It's doing DC work, and if the RAM was really marginal, though, I would expect to see errors or appcrashes or something.

 

[TheELF]

I don't know what to tell you man. I have a 2700 + X470 Taichi just like you. I upgraded from a 2500K system, just like you (although mine was clocked a bit higher). I'm currently sitting at 4ghz stable on all cores, at 1.35V in Bios and LLC5 (lowest level). Chip is in the low 70s running 16 threads of Prime 95 small FFT max heat. This system feels a LOT faster that my 2500k did. I assume some of that is going from a 1TB SATA SSD to a 1TB NVME SSD. Maybe try overclocking the system using a fixed speed? It's not like it's consuming a ton of power at idle. HW monitor shows values <10w. Package draw running prime is around 110-115W

I haven't tried using PBO/XFR or any of the auto overclocking features of the board yet. Maybe that'll come later, but I figured I'd start out with an old school OC to gather some information about the chip frequencies / voltages first.

He already said it multiple times,he doesn't want to overclock it to get the performance he was suggested that he would get.
You even quoted it.

It should be 4.1GHz boost, not 3.45GHz under lighter loads...and under heavy loads it shouldn't be 3.2GHz, it should be like 3.8GHz.