AMD Ryzen CPU's - Post your botting stats here

[Aion]

@Aion nice build but that two xeons are 260W. lots of heat to cool and power to consume wheres i7-7700K only 91W. How many you run on it?

The cooling inside Z800 is very good and professional , even on air, its decent and not being loud, so far. The extra power consume is not issue, those 150W max power correspond to 10-15 euro monthly in most EU countries, in US its even cheaper!

 

[Jiniix]

Actually, there is tons of variants, when building powerful botting PC from the aftermarket, the E5-2683 V3 is not the only option.
Im typing from my "brand new" Z800 Workstation, powered by 2x Xeon X5680 CPUs @ 3.3 GHz, paired with nVidia GTX1060 6GB and 32GB of DDR3 RAM.
And it is dirty cheap - $300 for the Z800 barebone, 2x CPUs $110 each, $180 for VGA, 60$ for RAM and $200 for the best SSD, available for this old mainboard.
Total $950 for fully working workstation with 12 cores (24 threads) @3.3GHz default / 3.6GHz boost. And its CPU benchmarks put it slightly better than brand new i7-7700K (4 cores) and i7-6800K with 6 cores.
Of course there is better performance rigs - its 30% slower than single i7-6900K (8 cores) but its worth $900 on Ebay only the CPU, without all the other stuff ...

Not quite sure what you're getting at. I'm saying it's a good idea to utilize used server hardware, but comparing direct performance between unrelated products are daft.
Using the 2683 as an example, you can buy one of those for $350 and a brand spanking new motherboard (incl. 20% VAT and an extra 20% tax from my country) for $220, giving you 14c/28t @ 2.0-3.0GHz at less than half the power usage (from CPUs isolated of course) and a more stable overall setup.
PS. $200 for an SSD is way overkill, unless you need a lot of storage for individual install folders or VMs.

 

[Aion]

Not quite sure what you're getting at. I'm saying it's a good idea to utilize used server hardware, but comparing direct performance between unrelated products are daft.
Using the 2683 as an example, you can buy one of those for $350 and a brand spanking new motherboard (incl. 20% VAT and an extra 20% tax from my country) for $220, giving you 14c/28t @ 2.0-3.0GHz at less than half the power usage (from CPUs isolated of course) and a more stable overall setup.
PS. $200 for an SSD is way overkill, unless you need a lot of storage for individual install folders or VMs.

But my point was to illustrate your post with details, I had recent experience with.

As a side note for the E5-2683, stay away of low-clocked many cores, since they do not tend to use the boost 3.0GHz on full load. I cannot say from personal experience with exact numbers, but as being told from a hardware technician, who build custom servers and workstations for the living this 2.0GHz default / 3.0GHz boost clock sample use its boosted speed only if one or a few cores are at 100% load, but once all 14 cores got on full load - in real-life scenario or in middle of benchmark test, their boost speed never reach the advertised 3.0GHz.
So it is better to get a Xeon with higher default clock, like the E5-2650 or so.
That was the reason why for my budged under $1000 I got after the old LGA1366 Xeon setup, due to their affordable X5680s, running at 3.3Ghz per 6 cores for only $110 per CPU.

I noticed this scenario even with it - Xeon X5680, default 3.3GHz, boost 3.6GHz, on full load in benchmark boost up to 3.43GHz, not a penny higher!


Btw, $200 SSD was for not regular handy-capped from the SATA3 interface @500MB/s SSD but for PCIe Kingston HyperX Predator 256GB 1400/1000 MB/s read/write. Yes it was not mandatory for botting rig but I just liked its speeds

For the fun, I even bought off ebay a boxed GTX1060, with box, signed by tons of Blizzard developers from the recent Blizzcon! It is better, when we pair the work with the fun after all

 

[brainAbuddy]

Ryzen is garbage. Get e5-2683 v3 for 350$ instead, 14 cores 28 threads. Ryzen is trash the 500$ 1080x loses to 350$ 7700k in every single game by a lot, for botting needs xeons are superior.

umm xeon is not specially better. they are way more expensive and yes you will get more cores, but that doesn't mean it's always better

 

[Jiniix]

As a side note for the E5-2683, stay away of low-clocked many cores, since they do not tend to use the boost 3.0GHz on full load. I cannot say from personal experience with exact numbers, but as being told from a hardware technician, who build custom servers and workstations for the living this 2.0GHz default / 3.0GHz boost clock sample use its boosted speed only if one or a few cores are at 100% load, but once all 14 cores got on full load - in real-life scenario or in middle of benchmark test, their boost speed never reach the advertised 3.0GHz

As a hardware technician, who builds custom servers and workstations (as well as high end gaming rigs), you are sorta correct. Turbo Boost involves a chart, detailing how many cores can boost to certain speeds based on workload. However, most consumer motherboards, which I used in my example, pretty much all include "Enhanced Turbo Boost", or whatever the manufacturer wants to call it, which makes all cores go to max boost speed, even if the workload only demanded one.
You may trade off ECC, direct PCI passthrough etc with those motherboards though.

 

[fellowbotter]

@Aion thats true E5-2683 has 3.0 Ghz Max Turbo Frequency for a SINGLE core. Per ALL cores is 2.5 Ghz I believe

 

[Druid07]

umm xeon is not specially better. they are way more expensive and yes you will get more cores, but that doesn't mean it's always better

Yes it is better for botting because more cores/threads means you can run more wow/hb/vmw instances. And it's also cheap if you buy used, like I said 350$ for the 14core 2683v3 or 90$ for 8core 2670, only idiots buy new xeons for personal use. For gaming there's 7700k. Ryzen = useless garbage

 

[fellowbotter]

Yes it is better for botting because more cores/threads means you can run more wow/hb/vmw instances. And it's also cheap if you buy used, like I said 350$ for the 14core 2683v3 or 90$ for 8core 2670, only idiots buy new xeons for personal use. For gaming there's 7700k. Ryzen = useless garbage

Yeah i7-7700K is good for gaming. Gaming needs less cores but higher clock speeds thats what i7-7700K got. Multibotting needs cores/threads and "some" clock speeds...

(Cinebench R15 multithreaded score in bold at the end)

Xeon E5-2683 v3 14c/28t (2.0-3.0) 1600
Ryzen 7 1700 8c/16t (3.0-3.7 ghz) 1400
i7-7700k 4c/8t (4.0-4.5 ghz) 900

If we only compare cinebench scores looks like that xeon wins over ryzen 7 1700 but just by a little and i7-7700k is way behind. I don't think Ryzen is garbage. For gaming yes it looses to i7-7700K but compared to that xeon it can be a powerfull botting cpu. Thats why I crated this topic. Still waiting for someone to test Ryzens out. Also if we compare cost. That xeon (used) + new mobo is around $550. For ryzen its $460 with mobo (both new). Not to mention 120W TDP and 65W.

 

[Druid07]

If we only compare cinebench scores looks like that xeon wins over ryzen 7 1700 but just by a little and i7-7700k is way behind. I don't think Ryzen is garbage. For gaming yes it looses to i7-7700K but compared to that xeon it can be a powerfull botting cpu. Thats why I crated this topic. Still waiting for someone to test Ryzens out. Also if we compare cost. That xeon (used) + new mobo is around $550. For ryzen its $460 with mobo (both new). Not to mention 120W TDP and 65W.

???
Are you gonna be botting Cinebench? Or maybe playing Cinebench? How is that particular benchmark relevant to my point? And then you've probably heard about Windows-Ryzen bug which gives Ryzen better results in Cinebench?
Let's check Corona Renderer benchmarks then:
Benchmark : Corona Renderer
You can see single 2683v3 in 500's and dual 2683v3 in 200's
Ryzen 1800x is in 1200's and Ryzen 1700 in 1600's. Bad!

Should be obvious anyways that you'll be able to run more hb/wow instances on 28 threads rather than just 16. Plus you can get dual socket mobo and go for 56 threads or even more with diff xeon's for exceptional botting/rendering performance. There's nothing exceptional in Ryzen.

 

[Jiniix]

And then you've probably heard about Windows-Ryzen bug which gives Ryzen better results in Cinebench?

Should be obvious anyways that you'll be able to run more hb/wow instances on 28 threads rather than just 16. Plus you can get dual socket mobo and go for 56 threads or even more with diff xeon's for exceptional botting/rendering performance. There's nothing exceptional in Ryzen.

How much is Intel paying you? Holy fuckballs.
"Ryzen bug" increasing Cinebench scores? Only Ryzen bug right now is core parking on Balanced power scheme, which decreases performance.
Your reasoning is seriously flawed. In two years, you'll buy 32c/64t Ryzen server CPUs used for the same price (and yes, you can use two at a time as well).

While you can get more performance now with used Intel server grade hardware, using that as a reason to call Ryzen garbage is nothing short of retarded.

Using your example, I can buy a 16 core AMD server CPU for $25 (supports up to four CPUs in one motherboard), which has waaaaaaaay more cores than Intels 4c/8t, therefore Intel is garbage.

 

[Druid07]

How much is Intel paying you? Holy fuckballs.
"Ryzen bug" increasing Cinebench scores? Only Ryzen bug right now is core parking on Balanced power scheme, which decreases performance.

Ryzen Platform Affected by RTC Bias; W8/8.1/10 Not Allowed on Select Benchmarks

In a statement issued by the Head of Moderation Christian Ney, we have confirmed that the AM4 platform is affected by the Windows 8/8.1/10 RTC bias. The bias occurs when adjusting the reference clock at run-time and will affect the Windows timer, causing benchmarks to perceive time slower (or faster) than it really is. This results in benchmark scores affected in such a way that the benchmark scores reported do not reflect real performance.

In two years...
While you can get more performance now...

Please. Is that the only argument AMD fanboys have now? "Just wait, Ryzen perfomance is crap because Windows needs optimization. And games don't use all the 16 threads, so just wait... Oh and there's also Naples coming which will end Intel's monopoly on the server market (Lisa Su told me so)"

Sorry, I don't take words of AMD public relations agents for granted. Show me actual benchmarks where Ryzen is superior to Intel in either gaming or server CPU's. Oh wait there are none.

Meanwhile you can get a 14core stepping 2 e5-2683 v3 for 350$ and a x99 mobo for 200$, unlock turboboost through bios hack on all the cores to 3ghz and get 2400 score on R15 Cinebench. That's just 550$ for cpu+mobo. And it's also 24/7 stable 50c max on air cooling so your 550$ Ryzen 1800x "world records" on liquid nitrogen are irrelevant.
Самый быстрый процессор Xeon E5-2683 v3 QS с разлоченным бустом (vs AMD Ryzen 1700X в Cinebech R15) - YouTube
1:12

 

[Jiniix]

Meanwhile you can get a 14core stepping 2 e5-2683 v3 for 350$ and a x99 mobo for 200$, unlock turboboost through bios hack on all the cores to 3ghz and get 2400 score on R15 Cinebench. That's just 550$ for cpu+mobo.
1:12

Literally what I said many posts ago. It's not even a BIOS hack, it's a common setting...
Read just a little bit, and you'd be surprised with the amount of available information :O

 

[Druid07]

Literally what I said many posts ago. It's not even a BIOS hack, it's a common setting...
Read just a little bit, and you'd be surprised with the amount of available information :O

You can't use CPU multiplier with that xeon, only after the bios hack.

 

[Aion]

You can't use CPU multiplier with that xeon, only after the bios hack.

But there is tons of Xeons, which run on default clock of 3.3-3.6GHz, with boost speed of 3.6-4GHz, which do not require bios hack to be performance kings in full load.

So while its good way for performance, its not the only one.

 

[Druid07]

But there is tons of Xeons, which run on default clock of 3.3-3.6GHz, with boost speed of 3.6-4GHz, which do not require bios hack to be performance kings in full load.

So while its good way for performance, its not the only one.

Yes, that Xeons have either few cores or cost like a booeing

 

[Jiniix]

You can't use CPU multiplier with that xeon, only after the bios hack.

You absolutely can. Not with a server board, but with just about any con/prosumer board - as I stated earlier.

 

[fellowbotter]

bump. Anyone tried Ryzen yet?

 

[schae]

Guess arround 12 bots in virtual machines, due to the ram limit of 32 gb. otherwise about 20. The 5960X/6900K can handle about 22-25 bots with WoD, in Legion its less, didnt reach the limit yet but it should be arround 16-18.

 

[cveks]

anyone tried this CPU for multiboting?