Would it make any sense, to upgrade Ryzen 1600 to 2600 (2nd-gen)?

Aug 25, 2001
42,469
236
126
#1
What if I suspect that my 1st-gen Ryzen 6C/12T CPU has the "Ryzen bug"?

I guess, it would get me +300-400Mhz? 3.8Ghz to 4.2Ghz?

The Ryzen 2nd-gen, are just a 12nm "shrink" of the 14nm Ryzen 1st-gen, right? Same masks for the most part? So no real arch. differences?

I figure that it would be pretty-much pointless, but I thought that I'd throw it out there.

What about a 1700X 8C/16T CPU? If they drop down to $149.99 New again?
 

epsilon84

Senior member
Aug 29, 2010
931
104
136
#2
What 'Ryzen bug' are you talking about exactly?

Zen+ has approx 3% higher IPC and 200MHz higher clockspeed ceiling compared to Zen (4.2GHz vs 4.0GHz)

Not really worth the upgrade from a 1600 to 2600 IMO. In a best case scenario you'll look at a 10% uplift, in many cases it will be closer to 5%.

The 1700X on the other hand is a more meaningful upgrade if you can make use of the extra cores.
 

Yuriman

Diamond Member
Jun 25, 2004
5,529
2
106
#3
I'm still on Ivy Bridge - you can guess my opinion. It used to be that I'd get a new CPU every time performance doubled, which was every 2-3 years. Now I'd do it for anything more than 25%. I have zero use for the extra cores, and per-core performance is only barely 25% better after 6 years. I use Coffee Lake CPUs every day at work and subjectively I can't tell a difference in every day tasks. 3% is completely meaningless.
 

whm1974

Diamond Member
Jul 24, 2016
6,091
203
96
#4
I'm still on Ivy Bridge - you can guess my opinion. It used to be that I'd get a new CPU every time performance doubled, which was every 2-3 years. Now I'd do it for anything more than 25%. I have zero use for the extra cores, and per-core performance is only barely 25% better after 6 years. I use Coffee Lake CPUs every day at work and subjectively I can't tell a difference in every day tasks. 3% is completely meaningless.
My guess is that you will need to wait until at least 2020 before seeing any major improvements in both CPU performance and features that it becomes worthwhile to build a new system.
 

Heclone

Junior Member
Dec 7, 2018
14
8
36
#5
I would recommend to wait until Zen2 R5 CPUs that should grants you a real boost.
Based on what we've seen with Zen+ (+ 3% IPC, + 7% Freq) and on what Zen2 should at least legitimately provides (+10% IPC, +10% Freq), you are probably looking for ~33% improved ST performance (and maybe two more cores) by jumping from Zen to Zen2 instead of only ~10% with the same core count.
 

epsilon84

Senior member
Aug 29, 2010
931
104
136
#6
I'm still on Ivy Bridge - you can guess my opinion. It used to be that I'd get a new CPU every time performance doubled, which was every 2-3 years. Now I'd do it for anything more than 25%. I have zero use for the extra cores, and per-core performance is only barely 25% better after 6 years. I use Coffee Lake CPUs every day at work and subjectively I can't tell a difference in every day tasks. 3% is completely meaningless.
I used to own a 3770K and now a 8700K, so I can agree that 'general desktop usage' hasn't been CPU limited for a long long time. I'm afraid if you're waiting for per core performance to double you'll probably be waiting for another decade. Thankfully software is getting increasingly multi-threaded, but yeah we are long past the point where a CPU upgrade brings a noticeable improvement to desktop usage. This isn't the 386 -> 486 -> Pentium era anymore ;)
 

Despoiler

Golden Member
Nov 10, 2007
1,799
42
136
#7
Personally I don't think it's worth it. I have a 1700 and thought about a 2700x. I could get up to 10% perf improvement in certain games. Mostly less I believe. I'm confident that Ryzen 3000 is going to launch by May/June of next year at the latest. Save your money and upgrade into something that will be crazy good.
 

Mopetar

Diamond Member
Jan 31, 2011
4,293
245
126
#8
I’d wait for the next Ryzen release since it’s still socket compatible. You’ll end up with more performance or if nothing else it will drive down the prices on existing chips and you can pick up a better deal.

If the leaks are remotely true, equivalent performance should get a hell of a lot cheaper.
 
Oct 9, 1999
11,181
29
126
#9
Im waiting out at least zen2 before i upgrade my 1700.
 

scannall

Golden Member
Jan 1, 2012
1,284
105
136
#10
Another vote for 'Wait'. The 1600 is a nice part. Allow yourself to be happy. ;-) It's OK, it really is.
 
Jan 29, 2014
267
5
91
#11
I 'upgraded' from an early R7 1700 (lost the silicon lottery big time) to a R7 2700x. It made a huge difference in some of the work I do (moving from 3.7ghz all core to 4.0 ghz all core + stable 3200mhz CL14) but made almost no difference in games. It depends on what you're doing. At this point I'd wait until the 3000 series either way.
 
Last edited:
Aug 25, 2001
42,469
236
126
#12
Another vote for 'Wait'. The 1600 is a nice part. Allow yourself to be happy. ;-) It's OK, it really is.
Yeah. Maybe I should just "enjoy". I'm a little neurotic about constant upgrades... and a little neurotic about BSODs. (Been getting them lately. Yes, I'm overclocked, on AIO 120mm water. 76C max.)
 

CHADBOGA

Golden Member
Mar 31, 2009
1,762
21
126
#13
Yeah. Maybe I should just "enjoy". I'm a little neurotic about constant upgrades... and a little neurotic about BSODs. (Been getting them lately. Yes, I'm overclocked, on AIO 120mm water. 76C max.)
Go down 100Mhz. o_O
 

coercitiv

Platinum Member
Jan 24, 2014
2,962
254
136
#15
and a little neurotic about BSODs. (Been getting them lately.
Already did. I'm at 3.80Ghz now, instead of 3.90Ghz.
This is why the extra price for the X chip tends to pay off in time. My 1600X has been running 24/7 for over a year now, rock solid system with 3.7Ghz all-core boost and 3.9-4.1Ghz low threaded boost. If I had to buy / recommend a CPU right now, I would point towards the 2600X over 2600 in a heartbeat. The only people who should opt for the 2600 are those with really stretched budgets.

This reminds me of a similar experience we had with Skylake, albeit that was even more exacerbated by more aggressive chip binning: IIRC @VirtualLarry bought and overclocked the i5 6400 (or was it the 7400?), then reported back with somewhat worrying figures on temps and voltage needed to go over 4Ghz. IIRC he decided to drop clocks there as well. Meanwhile I was sitting on a 6600K with virtual no OC (just 100Mhz to make it an even 4GHz @ 1.150V), with headroom to go for either a decent undervolt or a 15%+ overclock before voltage became a problem.

Chip quality offers a lot of flexibility that may be enough to carry you for 1-2 years more over the cheaper silicon. You can always win the lottery with a golden budget chip, but with the way technology works these days, odds are stacked up against you more and more.
 
Aug 25, 2001
42,469
236
126
#16
This reminds me of a similar experience we had with Skylake, albeit that was even more exacerbated by more aggressive chip binning: IIRC @VirtualLarry bought and overclocked the i5 6400 (or was it the 7400?), then reported back with somewhat worrying figures on temps and voltage needed to go over 4Ghz. IIRC he decided to drop clocks there as well. Meanwhile I was sitting on a 6600K with virtual no OC (just 100Mhz to make it an even 4GHz @ 1.150V), with headroom to go for either a decent undervolt or a 15%+ overclock before voltage became a problem.
That wasn't JUST a binning issue. BCLK OC on Skylake, overclocks not just the core clock, but also the Uncore and L3 cache. So while you were sitting pretty doing a multiplier overclock, while your cache/uncore sat at a lower, safer clock speed, my locked Skylake quad was pushing for the moon on all fronts, and the L3 cache was having trouble over 4Ghz.
 

coercitiv

Platinum Member
Jan 24, 2014
2,962
254
136
#17
That wasn't JUST a binning issue. BCLK OC on Skylake, overclocks not just the core clock, but also the Uncore and L3 cache. So while you were sitting pretty doing a multiplier overclock, while your cache/uncore sat at a lower, safer clock speed, my locked Skylake quad was pushing for the moon on all fronts, and the L3 cache was having trouble over 4Ghz.
My point was there's a price to pay for choosing the lower clocked chips, and if I end up with 1600X and 3600X while you choose 1600, 2600 and 3600 in a row because you actually want that top performance, it seems to me I get more value while sitting pretty.

PS: IIRC my uncore multi while doing oc was 42X, I might have gone higher (43-44) but my target was to find the upper clock limit.
 

Topweasel

Diamond Member
Oct 19, 2000
4,563
185
126
#18
I think we will find that long term over-clocks are going to kill Ryzen chips quicker than they ever did on Intel. Clockspeeds even as far back as Sandy Bridge seemed to be more of a thermal limit thing than a process limit. I think riding Ryzen near the process limit is going to cause more early deaths. Just looking to the first couple of weeks of Ryzen's releases we had people tringing to pump 1.4-1.45 volts to hit an overclock and realizing afterwards that now their CPU's couldn't even run safer overclocks with safer voltage any more. I don't know if that is what is happening with your CPU, but I wouldn't be surprised.
 
Aug 25, 2001
42,469
236
126
#19
I think we will find that long term over-clocks are going to kill Ryzen chips quicker than they ever did on Intel. Clockspeeds even as far back as Sandy Bridge seemed to be more of a thermal limit thing than a process limit. I think riding Ryzen near the process limit is going to cause more early deaths. Just looking to the first couple of weeks of Ryzen's releases we had people tringing to pump 1.4-1.45 volts to hit an overclock and realizing afterwards that now their CPU's couldn't even run safer overclocks with safer voltage any more. I don't know if that is what is happening with your CPU, but I wouldn't be surprised.
Sadly, I basically agree.

Although I never really pushed much beyond 1.4V.

Edit:
Here's a recent post that claims Ryzen degradation too.
https://forums.anandtech.com/thread...stupid-or-greedy.2521289/page-6#post-39674982
 
Last edited:

bfun_x1

Senior member
May 29, 2015
411
13
86
#20
I upgraded from a 1600 to a 2600X. Was the upgrade worth it? No. Would I do it again? Probably.
 

Heclone

Junior Member
Dec 7, 2018
14
8
36
#21
I upgraded from a 1600 to a 2600X. Was the upgrade worth it? No. Would I do it again? Probably.
Stock vs stock you do are getting some noticeable improvements : 3.2 - 3.7GHz to 3.6 - 4.2 GHz in addition of a small IPC improvement and a better memory support.

It is not a big jump, but you do get a real upgrade.
 

Elfear

Diamond Member
May 30, 2004
6,789
17
106
#22
I upgraded from a 1700 to a 2700X mainly due to the small upgrade cost ($15). I do notice a small difference in everyday tasks and gaming stock vs stock. The 1700 was kind of a pain to overclock and so I really enjoyed the great out-of-the-box performance of the 2700X.

That being said, if the cost to me wasn't so small, I wouldn't have upgraded.
 

bfun_x1

Senior member
May 29, 2015
411
13
86
#23
Stock vs stock you do are getting some noticeable improvements : 3.2 - 3.7GHz to 3.6 - 4.2 GHz in addition of a small IPC improvement and a better memory support.

It is not a big jump, but you do get a real upgrade.
The 1600 was at 3.8GHz on all cores. I left the 2600X stock so the biggest difference is the 4.2 GHz single core clock. Someday I'll probably swap it for a 3600X.
 
Aug 25, 2001
42,469
236
126
#24
It's weird, I just updated the firmware on my Ryzen R5 1600 rig, and as part of that, I reset to UEFI defaults, reboot, flash UEFI, allow it to reboot itself twice, enter UEFI, reset to defaults (new defaults, if any), reboot, enter UEFI, set XMP and RAM speed, reboot, optionally, enter UEFI and set overclock, reboot, and allow it to boot into Windows 10.

Well, I chose not to overclock it this time, and it seems... snappier? I don't quite understand it. Something about manual overclocking, seems to muck something up, regarding latencies somehow. Or maybe, it has something to do with the power plan, interacting with the scheduler, and whatnot, when manually overclocking. I don't know. Sure, some benchmarks are higher overclocked, but ... I like "snappy".
 

Heclone

Junior Member
Dec 7, 2018
14
8
36
#25
The 1600 was at 3.8GHz on all cores. I left the 2600X stock so the biggest difference is the 4.2 GHz single core clock. Someday I'll probably swap it for a 3600X.
3.8GHz on all core for a Zen1 non-X CPU is pretty good for h24 ! Didn't you try to overclock the 2600X ? However the perfomance gap with a 3600X will clearly be bigger.
 


ASK THE COMMUNITY