lynux3 2,461 Posted April 20, 2019 Share Posted April 20, 2019 Just now, The Mother Fucker said: No, so far I only tweaked my memory with it to get the true 3000Mhz rate speed and not the BS 2933Mhz it was getting. Hind sight, I probably should have got 3200 rate DDR4. I'll be interested to see the results in game mode. Link to post Share on other sites
JonDnD 3,032 Posted April 20, 2019 Share Posted April 20, 2019 computers Link to post Share on other sites
The Mother Fucker 27 Posted April 20, 2019 Author Share Posted April 20, 2019 6 minutes ago, Nya` said: Here's when I was running two 290X's. That's interesting to see. I wonder what a single RX 290x score would look like. Link to post Share on other sites
Remij 5,048 Posted April 20, 2019 Share Posted April 20, 2019 3 hours ago, lynux3 said: I'll be interested to see the results in game mode. Game mode will almost certainly make things perform worse in 3dmark. It's meant to scale with cores. Link to post Share on other sites
lynux3 2,461 Posted April 20, 2019 Share Posted April 20, 2019 42 minutes ago, Remij_ said: Game mode will almost certainly make things perform worse in 3dmark. It's meant to scale with cores. I don't think you understand how Threadripper works. 1 Link to post Share on other sites
The Mother Fucker 27 Posted April 20, 2019 Author Share Posted April 20, 2019 38 minutes ago, Remij_ said: Game mode will almost certainly make things perform worse in 3dmark. It's meant to scale with cores. I dunno about that. If the performance regression theory is to be believed which I really think it is true. (see my next post) Disabling dies will increase the performance. https://techgage.com/article/threadripper-2990wx-performance-coreprio/ Now I've used Coreprio before but didn't have good results and it was only as an alternative because AMD has the Dynamic Local Mode feature in Ryzen Master that does the job, and previously I could not run Ryzen Master because I was using Hyper-V. Link to post Share on other sites
The Mother Fucker 27 Posted April 20, 2019 Author Share Posted April 20, 2019 Okay, guys, I said when I got back home (from work) I would play around with more profiles in Ryzen Master. First course was to try to get my 2990wx TR to run like a 2950x TR with just 16 cores and 32 threads. Mind you this is not even the "Game Mode' feature yet of Ryzen Master, which will disable all 3 cores, leaving just 1/4. Which is like a Ryzen 7 2700x. Here are my new Fire Strike score. Only difference here is me disabling half the CPU dies of the 32-core Threadripper. 1 Link to post Share on other sites
The Mother Fucker 27 Posted April 20, 2019 Author Share Posted April 20, 2019 Link to post Share on other sites
lynux3 2,461 Posted April 20, 2019 Share Posted April 20, 2019 Threadripper is a fuckin beast, but I don't know what I'd do with it. My 2700X does everything I need right now. Link to post Share on other sites
lynux3 2,461 Posted April 20, 2019 Share Posted April 20, 2019 In terms of HEDT the Threadripper shits all over everything in price and performance. lmao @ Infail Link to post Share on other sites
Remij 5,048 Posted April 20, 2019 Share Posted April 20, 2019 3 hours ago, lynux3 said: I don't think you understand how Threadripper works. I think I understand exactly. 3dmarks physics test scales with cores dumbass. It will almost certainly lower his physics score... @The Mother Fucker Now enable all the cores, and run using only game mode and post the results. I want to shut this dumb bitch up Link to post Share on other sites
Remij 5,048 Posted April 20, 2019 Share Posted April 20, 2019 3 hours ago, The Mother Fucker said: Mind you this is not even the "Game Mode' feature yet of Ryzen Master, which will disable all 3 cores, leaving just 1/4. Which is like a Ryzen 7 2700x. Exactly... so why the fuck are you guys saying I'm wrong that enabling game mode will reduce performance in 3dmark? lmao.. Link to post Share on other sites
The Mother Fucker 27 Posted April 20, 2019 Author Share Posted April 20, 2019 Results summary of Fire Strike benchmark with the 32-core / 64 thread AMD Ryzen Threadripper 2990WX. 1 die - 2700X Quote Here is the Threadripper in the "GAME MODE" setting of Ryzen Master. Operating on just 1/4 die cores, think of Ryzen 7 2700X 2 dies - 2950X Quote Here is the Threadripper operating in half-off, AMD Ryzen Master utility refers as Legacy Compatibility Mode (a stupid name because it's actually better than GAME MODE) This is 1/2 of the dies enabled, think of the Ryzen Threadripper 2950X 4 dies - 2990wx Quote Last here is the Threadripper in full 4 die operation, 32-cores / 64 threads , two of the cores having direct memory access thus performance regression. Link to post Share on other sites
The Mother Fucker 27 Posted April 20, 2019 Author Share Posted April 20, 2019 (edited) 5 hours ago, Remij_ said: Exactly... so why the fuck are you guys saying I'm wrong that enabling game mode will reduce performance in 3dmark? lmao.. I guess what I was trying to say is that if I were to disable dies on my 2990wx Threadripper it would improve it's 3DMark performance. Ryzen Master's "Game Mode" is a preset mode that disables 3 of the 4 dies of the 2990wx which is not what I want. If you looked at the die diagram shown in the link I posted earlier, I'll post it here now: I knew from that disabling half the dies of my Threadripper would present better performance. As 2 of the dies have direct access to the local RAM. Clearly dropping the 2990wx to just a single die would be reduced performance. As the results I just posted earlier show. Edited April 20, 2019 by The Mother Fucker 1 Link to post Share on other sites
DynamiteCop 2,161 Posted April 20, 2019 Share Posted April 20, 2019 2 minutes ago, The Mother Fucker said: Results summary of Fire Strike benchmark with the 32-core / 64 thread AMD Ryzen Threadripper 2990WX. 1 die - 2700X 2 dies - 2950X 4 dies - 2990wx This is pretty interesting. Link to post Share on other sites
Remij 5,048 Posted April 20, 2019 Share Posted April 20, 2019 @The Mother Fucker @Nya` Yea it's quite interesting stuff. What I'm wondering though is why the Graphics score takes a massive hit under the default setting? Can you confirm The Mother Fucker that both GPUs are being utilized.. or rather how much they are? Because in both other configurations they are being fully used. I assumed (pretty strongly) that only 1 GPU was being utilized.. but it very well could be that both are and there's simply a bug with TR and Crossfire above 2 dies, which impacts utilization. (That would make sense since the other Threadripper CPUs I checked didn't exhibit this behavior with Crossfire. Anyway.. what a beastly machine. Is that your own personal rig or do you use it for work as well? Link to post Share on other sites
kaz 2,493 Posted April 20, 2019 Share Posted April 20, 2019 he uses it at work to type in ms word Link to post Share on other sites
The Mother Fucker 27 Posted April 20, 2019 Author Share Posted April 20, 2019 1 hour ago, Remij_ said: @The Mother Fucker @Nya` Yea it's quite interesting stuff. What I'm wondering though is why the Graphics score takes a massive hit under the default setting? Can you confirm The Mother Fucker that both GPUs are being utilized.. or rather how much they are? Because in both other configurations they are being fully used. I assumed (pretty strongly) that only 1 GPU was being utilized.. but it very well could be that both are and there's simply a bug with TR and Crossfire above 2 dies. (That would make sense since the other Threadripper CPUs I checked didn't exhibit this behavior. Anyway.. what a beastly machine. Is that your own personal rig or do you use it for work as well? I built this system purely as an ultimate build project that I was going to test around cryptomining with. I have 3 other PCs, two of them are primary and have been undergoing upgrades, one of my primary system at the moment uses a Ryzen 5 2600, 32GB DDR4 with two 980Tis. It replaced a Phenom II X4 that used the GTX 560. The other soon to be built will be a FX 8300, 32GB DDR3 with two GTX 670s or single GTX 770. It will replace my aging Opteron 185 2.6GHz, 4GB DDR1 PC The 4th PC waiting on me to build will use a Ryzen 5 2400G, 32GB DDR4 Corsair Vengeance, and depending on how well the APU fair, I might add my GTX 1660 Ti to it. The 4th PC was excessive and likely the result of me watching too much PC hardware Youtubers and getting influenced to build. 1 Link to post Share on other sites
bhytre 2,815 Posted April 20, 2019 Share Posted April 20, 2019 Remipee's PosC BTFO Link to post Share on other sites
Remij 5,048 Posted April 20, 2019 Share Posted April 20, 2019 27 minutes ago, The Mother Fucker said: I built this system purely as an ultimate build project that I was going to test around cryptomining with. I have 3 other PCs, two of them are primary and have been undergoing upgrades, one of my primary system at the moment uses a Ryzen 5 2600, 32GB DDR4 with two 980Tis. It replaced a Phenom II X4 that used the GTX 560. The other soon to be built will be a FX 8300, 32GB DDR3 with two GTX 670s or single GTX 770. It will replace my aging Opteron 185 2.6GHz, 4GB DDR1 PC The 4th PC waiting on me to build will use a Ryzen 5 2400G, 32GB DDR4 Corsair Vengeance, and depending on how well the APU fair, I might add my GTX 1660 Ti to it. The 4th PC was excessive and likely the result of me watching too much PC hardware Youtubers and getting influenced to build. Nice. You're an AMD lad aren't you lol. I'm going to build a Ryzen 3000 based PC when those drop within the next 3-4 months. Ryzen 9 3850X 16 core 32 thread @ 4.5-5GHz New PCIe 4.0 based motherboard 32GB of extremely fast RAM Tons of storage I think I'm gonna go big on storage for my next build. 2TB of SSD + 4TB of HDD isn't cutting it anymore. I'd like to ditch HDDs completely. 1 Link to post Share on other sites
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now