Precision 7540 & Precision 7740 owner's thread

[Nrml Man]

Hello everybody 👋 Now i7-8750h and quadro p4200 8gb on my 7730.I saw p5200 16gb for a "tasty" price.Is it any sense to buy it (i mean will my cpu reveal this gpu?),and need i apply other heatsink with it?

[Aaron44126]

Because of the power cap, performance of P4200 and P5200 will be nearly identical in this system. The only reason to upgrade would be if you can make use of the extra VRAM.

[Nrml Man]

And what about rtx4000/5000? Will be necessary other heatsink or i can use default?

[Aaron44126]

Cards from the Precision 7740 can be used, but as is typical with this sort of cross-generation upgrade, you will need an INF mod to load the NVIDIA driver under Windows and that can mess with a small number of games that are particular about driver signatures for anti-cheat purposes.

 

I'm not sure about the heatsink. I don't think a swap is needed but I can't confirm. Carefully check photos of the cards and make sure that screw holes and VRMs are in the same position.

[Nrml Man]

Okay,thanks.By the way,i didn't have any problem with Quadro T2000 on my 7530 (Inf mod) playing Chivalry 2 (from Epic's store,it uses Easy Anti-Cheat).

[Annihilator]

10 hours ago, Aaron44126 said:

Because of the power cap, performance of P4200 and P5200 will be nearly identical in this system. The only reason to upgrade would be if you can make use of the extra VRAM.

 

14% difference i not would call it nearly same performance(both at 100W) but yeah, they r close, n the price difference is huge from P4200 to P5200, for the 14% n double mem u have to pay more then double price 😬 on the other hand, u can cut the TDP from the P5200 to 75W n get the same performance like the P4200 but much cooler n less noisy + double mem. 

 

n still 💩 never a P6000 mobile was out(maybe prototypes), a full GP102 at 100-120W with 24GB would be really nice 🤩

[Aaron44126]

1 hour ago, Annihilator said:

14% difference i not would call it nearly same performance(both at 100W)

 

Every benchmark that I saw when these launched put the difference at <5% if there was not VRAM pressure.

[Annihilator]

1 hour ago, Aaron44126 said:

 

Every benchmark that I saw when these launched put the difference at <5% if there was not VRAM pressure.

 

P5200 vs. P4200

 

[Aaron44126]

11 hours ago, Annihilator said:

P5200 vs. P4200

 

This seems a bit flaky to me. They don't give details of the system that these tests were run on. There are anomalies like P4200 actually winning their Vulkan test.

 

I'll admit to not digging into the tests here too much, because I already know how it's going to shake out. I've been closely following this line of systems for 15+ years and this is something that has come up for discussion and testing, over and over and over again. There was a lot of chatter about this on NotebookReview at the time of the launch of these systems. Users who had the systems in hand tried different things. P4200 and P5200 were barely different in most cases.

 

This really started to get a lot of attention with the Pascal generation as it started to become clear around that time how NVIDIA was putting very power-constrained desktop-class chips that could "use" a lot more power into laptops, and the power limit was the limiting factor on their performance. But the same has held true with every generation since then, comparing 4000- and 5000- series GPUs from NVIDIA in mobile workstations. It makes sense, given that the architecture and the power limit are the same. Spreading out the load over more cores is going to help with efficiency, but not to the tune of >10% performance. You're welcome to check yourself if you get one in hand.

[Annihilator]

4 hours ago, Aaron44126 said:

 

This seems a bit flaky to me. They don't give details of the system that these tests were run on. There are anomalies like P4200 actually winning their Vulkan test.

 

I'll admit to not digging into the tests here too much, because I already know how it's going to shake out. I've been closely following this line of systems for 15+ years and this is something that has come up for discussion and testing, over and over and over again. There was a lot of chatter about this on NotebookReview at the time of the launch of these systems. Users who had the systems in hand tried different things. P4200 and P5200 were barely different in most cases.

 

This really started to get a lot of attention with the Pascal generation as it started to become clear around that time how NVIDIA was putting very power-constrained desktop-class chips that could "use" a lot more power into laptops, and the power limit was the limiting factor on their performance. But the same has held true with every generation since then, comparing 4000- and 5000- series GPUs from NVIDIA in mobile workstations. It makes sense, given that the architecture and the power limit are the same. Spreading out the load over more cores is going to help with efficiency, but not to the tune of >10% performance. You're welcome to check yourself if you get one in hand.

 

 

i know, was getting worse with turing where some companies allowed 2070M a higher powerlimit then their 2080M ones n so the 2070 was outperforming the 2080. 😄

 

n u can compare desktop 1080 FE vs. 1070Ti FE(has slightly more shaders then the P4200, so will couple % xtra fps) both has 180W TDP... here 4K benchmarks(so no CPU limit everywhere, system on both cards: MSI Z270 Gaming Pro Carbon, i7-7700K, 2x 16GB DDR4 3200 C14), AVR/1% low FPS:

 

BF 1:  1080 73/69 - 1070Ti 66/61: +22/13%

Destiny 2: 1080 67/57 - 1070Ti 60/51: +12/12%

Metro Last Light Redux:  1080 66/42 - 1070Ti 60/40: +10/5%

Middle Earth Shadow War: 1080 60/46 - 1070Ti 54/42: +11/10%

RotTR: 1080 44/34 - 1070Ti 40/30: +10/13%

Tom Clancys Ghost Recon Wildlands: 1080 50/38 - 1070Ti 47/35: +6/9%

Tom Clancys The Division: 1080 53/42 - 1070Ti 50/39: +6/8%

Warhammer 40k Dawn of War III: 1080 52/48 - 1070Ti 48/44: +8/9%

Witcher 3: 1080 56/48 - 1070Ti 52/43: +8/12%

 

means average fps +10% n 1% low even +10%, n when u count down the xtra shaders the 1070Ti has compared to the P4200(which is between 1070 n 1070Ti) shaders will be around 12%. 

 

those low performance gain u mean, is pretty sure bcs of CPU limit(1080p tests), but in the GPU limit the P5200 will always clearly outperform the P4200 at same TDP limit. 

 

*%results rounded up/down