Threadripper Air Coolers-Silver Arrow TR4 vs. Wraith Ripper @ [H]

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,532
Threadripper Air Coolers-Silver Arrow TR4 vs. Wraith Ripper

Cooling the AMD Threadripper CPUs is not to be taken for granted. We pit the Thermalright Silver Arrow TR4 against the Cooler Master Wraith Ripper using both the Ryzen Threadripper 2990WX and 2950X processors. We show you how both these air coolers perform using Precision Boost 2 and Precision Boost Overdrive.

If you like our content, please support HardOCP on Patreon.

NOTICE: We have gone back and updated all graphs in this review with 100% fan speeds on both coolers. As noted previously, we used the default fan curve for the original version of this article. We decided that we needed to add the fans pushed to 100% on both coolers and further look at the data. We have not changed our written analysis.

Speaking in general terms for all sets of data here, we could realize about a ~650RPM increase on the Wraith Ripper (2560RPM max), which pushed decibel readings from ~46dB to ~48dB, which not much real change in sound profile at all. With the Silver Arrow, we saw up to 1100RPM changes (2350RPM max), which pushed decibels from ~46db to ~53dB, which greatly changed the sound profile. It was loud, but not exactly irritating.
 
Last edited:
Extremely informative review about such simple items as fans. SInce there are only two primary air coolers on the market for the TR4 socket you guys have the ultimate guide between the two which is going to help literally hundreds if not more existing threadripper customer.

However, that XSPC Raystorm you compared them to. If it werent for your review of the Raystorm long ago many would not even have considered it. I run that block on my 2950x and its an absolute cooling monster provided you give it ample radiator.

Thanks K and crew for another great and useful review.
 
Extremely informative review about such simple items as fans. SInce there are only two primary air coolers on the market for the TR4 socket you guys have the ultimate guide between the two which is going to help literally hundreds if not more existing threadripper customer.

However, that XSPC Raystorm you compared them to. If it werent for your review of the Raystorm long ago many would not even have considered it. I run that block on my 2950x and its an absolute cooling monster provided you give it ample radiator.

Thanks K and crew for another great and useful review.
Damn review was a rabbit hole.
 
I thought the clock speed difference was to be expected. But the power draw? Damn. And it's amazing to think that it's all doing that on its own without manual settings having to be specified. It both adds to and takes the fun out of OCing at the same time...

I still think it would be need to jury-rig a cooler that large into a socket it wasn't designed for.
 
I thought the clock speed difference was to be expected. But the power draw? Damn. And it's amazing to think that it's all doing that on its own without manual settings having to be specified. It both adds to and takes the fun out of OCing at the same time...

I still think it would be need to jury-rig a cooler that large into a socket it wasn't designed for.
 
Been eagerly waiting this review almost as much as I've been waiting for a Silver arrow TR4 to show up in stock somewhere. I want to swap my Bykski water block out for one so I don't have to worry about water cooling a computer I rarely look at. Hey Kyle, if you want to trade so you can have a Bykski to include in testing again lmk!
 
Been eagerly waiting this review almost as much as I've been waiting for a Silver arrow TR4 to show up in stock somewhere. I want to swap my Bykski water block out for one so I don't have to worry about water cooling a computer I rarely look at. Hey Kyle, if you want to trade so you can have a Bykski to include in testing again lmk!
Thanks. We still have the Bykski around. I gave it Dan, he uses it on his TR test bench.
 
Thanks for this review! I have been waiting since your teaser a while back. I am curious that you guys have not had the Noctua TR4 coolers on your test bench at all. Lots of us in the forums are using them since we didn't really have alot of other options while wanting to stay on air. Any plans to test NHU14S ? i would be interested to see how much head room swapping in the thermalright would give me since it looks a little bigger and is rated for higher TDP
 
As always learning a lot when reading these. The 2950x, or something similar, is under consideration for my next build in a year or two. Based on what your tests and conclusions have shown that seems to be the tier limit for what extra performance air can provide. Not to mention I like the 2950x price point. Don't think I've ever seen a split design with the fan in the middle before. Ultimately I'd probably pick the wraith ripper. This hypothetical build will go into the HAF X my 2600k resides in currently. That WR looks like it was totally designed with a HAF case in mind. I added a 2nd fan to the top last year so a lot of air flow.

Just from doing my own builds I totally understand there's a lot of details to consider, size, ability, costs, loudness, where your directing the air, and that's not even including the CPU features of these chips. Swapping out and examining all those metrics for 2 cpu's is a lot of work so thanks for the efforts with these.
 
Thanks for this review! I have been waiting since your teaser a while back. I am curious that you guys have not had the Noctua TR4 coolers on your test bench at all. Lots of us in the forums are using them since we didn't really have alot of other options while wanting to stay on air. Any plans to test NHU14S ? i would be interested to see how much head room swapping in the thermalright would give me since it looks a little bigger and is rated for higher TDP
I have three Noctua coolers here to review.
 
The rare coveted double gold.

I wonder if anyone has tried digging into PBO to possibly increase the 68C hard limit. Could be interesting if PBO was able to really stretch its legs up to higher temps.....or just fry the CPU
 
Well done, very impressed that they could cool a 2990wx and run it above base clock. Only 50mhz difference at stock between the silver arrow and the xspc setup is impressive - but if investing almost $2k in a cpu, I think I'd want the extra 500mhz that the xspc+pbo gives. Then again, maybe I'd be a little wary about burning up a $2k chip or having water dump on it....Good food for thought.
 
HMMMMM. Very Good article but... A cooler is only as good as its placement. Seen too many rigs over the year with bad air flow problems.

I am running my 1800X with a Cooler Master Hyper T2 with absolutely no problems what so ever with really low temp numbers. My whole rig runs cool. Air cooling only. This is one of my specialties. Old Skool testing to get maximum performance for the price you have to invest. But again good article :)
 
Thanks kyle for the indepth review of both coolers.

I kinda expected more of the SilverArrow TR4, given that it still won over the Wraith ripper, i kinda though it was going to be more temperature difference, but maybe it can be under the same fan and same rpms or normailizing the noise/db to see the outcome of the temperature.
 
HMMMMM. Very Good article but... A cooler is only as good as its placement. Seen too many rigs over the year with bad air flow problems.

I am running my 1800X with a Cooler Master Hyper T2 with absolutely no problems what so ever with really low temp numbers. My whole rig runs cool. Air cooling only. This is one of my specialties. Old Skool testing to get maximum performance for the price you have to invest. But again good article :)

Old skool? I'm sure more than a few ofd us old farts have done this: :D
duct.jpg
 
I would love to have one of these systems......but I would never ever use them to the extent they are made for. I just love coming here and seeing what these beasts can do (CPUs).

I really do love the no bullshit style and straight up facts when it comes to reviewing products on this site, keep up the good fight Kyle and everyone at the [H], thanks for the review as always!
 
I just sent my Enermax Liqtech back on account of all the people talking about leaks. Here's to hoping CoolerMaster is not lying about the first week of October for the availability of the Wraith Ripper!!! Thank you, Kyle, for the great review!!
 
NOTICE: We have gone back and updated all graphs in this review with 100% fan speeds on both coolers. As noted previously, we used the default fan curve for the original version of this article. We decided that we needed to add the fans pushed to 100% on both coolers and further look at the data. We have not changed our written analysis.

Speaking in general terms for all sets of data here, we could realize about a ~650RPM increase on the Wraith Ripper (2560RPM max), which pushed decibel readings from ~46dB to ~48dB, which not much real change in sound profile at all. With the Silver Arrow, we saw up to 1100RPM changes (2350RPM max), which pushed decibels from ~46db to ~53dB, which greatly changed the sound profile. It was loud, but not exactly irritating.
 
It would be nice if noctua made TR version of the nh-d15. I think it could handle the load fine
 
Thanks for the good work.

So Silver Arrow TR4 is 11c cooler at full speed and 53dB, but still 9.5c warmer than your XSPC Raystorm. What is the noise level of XSPC Raystorm system in these tests? Have your ran any test on 2990WX with Liqtech TR4 and if so what is it's temps and noise level?
The four fans at full RPM on that rad are 45.5db at 4 feet.

Given the quality problems shown on the Liqtech coolers, we have rescinded our awards and have not done any further testing. Those reviews are still up however.
 
Glad they are in stock!

Be sure you increase your case airflow to Silver Arrow TR4 because the TY-143 fan move twice as much air as normal fans do at full speed and if you do not increase your case airflow to cooler it will be re-using it's own heated exhaust air to make up the difference .. and every degree warmer the air is going into cooler translated to a degree hotter the CPU is. ;)

It's sad we can't buy Thernalright's TY-143 SQ fans in USA. Same fan as on cooler in a square housing. They work nicely with high performance cooler fans to supply needed airflow to cooler.
 
Kyle, just wondering, have you tested the cooler with the Gigabyte X399 Aorous Xtreme?

I'm thinking of getting that board but not 100% sure if the cooler will block the first PCI-e slot. It looks like on the Gigabyte board, there's a bit less space between the slot and where the CPU is compared to other X399 boards.

There's not a huge difference so I think it should be ok, but would be good to know for sure
 
  • Like
Reactions: doyll
like this
Kyle, just wondering, have you tested the cooler with the Gigabyte X399 Aorous Xtreme?

I'm thinking of getting that board but not 100% sure if the cooler will block the first PCI-e slot. It looks like on the Gigabyte board, there's a bit less space between the slot and where the CPU is compared to other X399 boards.

There's not a huge difference so I think it should be ok, but would be good to know for sure
Quick measurement....X399 Xtreme looks to have the first PCIe slot 5mm closer to the TR4 socket than the MSI MEG X399.
 
Kyle, just wondering, have you tested the cooler with the Gigabyte X399 Aorous Xtreme?

I'm thinking of getting that board but not 100% sure if the cooler will block the first PCI-e slot. It looks like on the Gigabyte board, there's a bit less space between the slot and where the CPU is compared to other X399 boards.

There's not a huge difference so I think it should be ok, but would be good to know for sure
Quick measuring gives approx 70.3mm center CPU to near side of x16 PCIe socket. Silver Arrow TR4 is 67.8mm center base to side of fins on short side and 86.8mm on long side.
 

Attachments

  • Silver Arrow TR4 meas. bst.png
    Silver Arrow TR4 meas. bst.png
    294.6 KB · Views: 0
1538050914h8ycl6a1j9_2_14_l.jpg


11mm on the MEG, so you should have 6mm on the Xtreme. I think I took a pretty accurate measurement yesterday morning. I measured the boards here in the office.
 
Important thing is it will fit without issues other than TY-143 fan moving twice as much are as any other cooler except for old HE01 with FHP141 fan. Both need case airflow increased by at least 75cfm more airflow than any system with coolers using normal speed/airflow fans.
 
Back
Top