ADVICE NEEDED! Black Friday Due date

[mpfiest]

So my boss told me that I should buy myself a new desktop computer for Black Friday, but told me yesterday… I’ve been non stop researching trying to get parts that are compatible and meet my needs.

 

I’ve never built a PC before. But I did build a parts list.

 

My two biggest needs are:

1. I want to confirm everything is compatible / what part changes are needed.

2. I have 0 clue when it comes to the wiring.. so really need help with that.

3. Below 2K

 

I’m in data science, handle large data models, will be progressing into machine learning eventually. The games I play are Steep, Cities Skylines, Skyrim, Riders Republic, and other open world RPGs.

Here are my parts:

 

SAMSUNG 980 PRO SSD with Heatsink 1TB PCIe Gen 4 NVMe M.2 Internal Solid State Drive + 2mo Adobe CC Photography, Heat Control, Max Speed, PS5 Compatible (MZ-V8P1T0CW) https://a.co/d/36BUrGC

 

6 Case Cooling fans

Noctua NF-P12 redux-1700 PWM, High Performance Cooling Fan, 4-Pin, 1700 RPM (120mm, Grey), Compatible with Desktop https://www.amazon.com/dp/B07CG2PGY6?psc=1&ref_=cm_sw_r_apin_ct_3BJCQN1W749D4VQ72D50&language=en_US

 

Case

Antec Performance 1 FT W, RTX 40 Series GPU Support, Temp. Display, 4 x Storm T3 PWM Fans, Type-C, Dual TG Side Panels, Removable Top Fan/Radiator Bracket, Mesh Front Panel, Full-Tower E-ATX PC Case https://www.amazon.com/dp/B0CH83WYVL?psc=1&ref_=cm_sw_r_apin_ct_TXE56K8JDFCFTQ4VZKF3&language=en_US

 

Adhesive

ARCTIC MX-6 (8 g) - Ultimate Performance Thermal Paste for CPU, Consoles, Graphics Cards, laptops, Very high Thermal Conductivity, Long Durability, Non-Conductive, CPU Thermal Paste https://www.amazon.com/dp/B09VDLH5M6?psc=1&ref_=cm_sw_r_apin_ct_A4DZA03AWKDGK85DM06B&language=en_US

 

Power Supply

Corsair RM1000x (2021) Fully Modular ATX Power Supply - 80 PLUS Gold - Low-Noise Fan - Zero RPM - Black https://www.amazon.com/dp/B08R5PH1VY?psc=1&ref_=cm_sw_r_apin_ct_RCZ92KC96ZT6DT4DY5DD&language=en_US

 

CPU cooling

Liquid Cooler NZXT Kraken 360 RGB - 360mm AIO CPU Liquid Cooler - Customizable 1.54" Square LCD Display for Images, Performance Metrics - High-Performance Pump - 3 x F120 RGB Core Fans - Black https://www.amazon.com/dp/B0BY3FD274?psc=1&ref_=cm_sw_r_apin_ct_3E0JZ7F4H9ZC89TCJ4Q9&language=en_US

 

GPU

XFX Speedster SWFT309 AMD Radeon RX 6700 XT CORE Gaming Graphics Card with 12GB GDDR6 HDMI 3xDP, AMD RDNA 2 RX-67XTYJFDV https://www.amazon.com/dp/B08YKCC8XD?psc=1&ref_=cm_sw_r_apin_ct_S4AK2QZ6MFG5M6Y9C6XH&language=en_US

 

CPU and motherboard

Intel Core i7-13700K (Latest Gen) Gaming Desktop Processor 16 cores (8 P-cores + 8 E-cores) with Integrated Graphics - Unlocked and ASUS Prime Z790-A WiFi 6E LGA 1700(Intel®13th&12th) https://www.amazon.com/dp/B0CF1VN2DT?psc=1&ref_=cm_sw_r_apin_ct_S749Z55JTFSNE7DE1GYE&language=en_US

 

RAM

Patriot Memory Viper Steel DDR4 64GB (2 x 32GB) 3600MHz Kit - PVS464G360C8K https://www.amazon.com/dp/B08688GFPD?psc=1&ref_=cm_sw_r_apin_ct_3FXTM5JGPY9KQNVTN8RD&language=en_US

 

Let me know your thoughts! Thank you!

[m1lk3y]

It's not a bad overall build, and I don't see any compatibility issues, although I have a few suggestions.

-You mentioned that you work with large data models but only have a 1TB SSD? I would definitely recommend getting more if you can afford it, storage is very cheap right now and I think having that extra space would save you headache in the future. 1TB doesn't last long these days, especially if you're going to be working and gaming on the same rig.

-I don't see the need for the Noctua fans. The case you picked already comes with 4 very nice fans, and then you have an additional 3 fans on the AIO you picked. So unless you were going to replace them all and then throw out the old fans, they aren't needed.

-Thermal paste is also not needed since the AIO will have pre applied thermal paste. Save purchasing your own tube for down the road when it comes time to clean and reapply.

 

If you decided to change anything, pcpartpicker has a great built in compatibility tool, and if you're ever confused about putting it together (wiring or anything), there are a million step by step videos on YouTube that can set you up in the right direction. Just don't watch the video from the Verge. Hope this helps!

 

edit: also welcome to the forum!

[RONOTHAN##]

Please use PCPartPicker.com for making the parts list, it does the initial compatibility checks (it's not perfect, there are some parts that it says are compatible that actually aren't, but it's a good first check), finds the cheapest parts available, and is also much easier to view than a slew of Amazon links.

 

Anyway, this is probably the system I'd actually build: https://pcpartpicker.com/list/PfD97R

1. The cooler is a bit cheaper for about the same performance. The thermal paste is unnecessary as CPU coolers come with some.
2. The motherboard is more feature rich, though admittedly will require doing a BIOS update to work (this isn't too difficult, look up BIOS Flashback on MSI motherboards if you want to see a tutorial on how to do so)
3. The RAM is DDR5 as a) it's faster, especially for workstation tasks and b) it's what the motherboard supports (the list you have would not work as you have a DDR5 motherboard and a DDR4 memory kit)
4. The SSD is higher capacity, ever so slightly faster, and currently Western Digital has a better reputation than Samsung (Samsung has been plagued with QC issues over the past year or so). 
5. The case is a bit cheaper, though still super easy to build in for a first time builder. The Noctua fans you have aren't necessary.
6. The PSU is plenty good enough, a 1000W unit is unnecessary.
7. The GPU is a 4070 Ti. If you're planning on doing deep learning and other types of machine learning algorithms, you really should be using a Nvidia card. Most of the libraries out there for doing this are designed around using Nvidia technologies such as CUDA and the Tensor cores, and while there are some that can use AMD cards like the 6700 XT you selected, they're kind of rare and the Nvidia cards are generally better supported in this field. That said, if you want to save a bit of money, dropping down to the regular 4070 wouldn't lose you too much machine learning performance, and realistically for the games you're playing it should still be plenty adequate. 

There is a white theme to it, though that was mostly by accident. 

[mpfiest]

23 minutes ago, m1lk3y said:

It's not a bad overall build, and I don't see any compatibility issues, although I have a few suggestions.

-You mentioned that you work with large data models but only have a 1TB SSD? I would definitely recommend getting more if you can afford it, storage is very cheap right now and I think having that extra space would save you headache in the future. 1TB doesn't last long these days, especially if you're going to be working and gaming on the same rig.

-I don't see the need for the Noctua fans. The case you picked already comes with 4 very nice fans, and then you have an additional 3 fans on the AIO you picked. So unless you were going to replace them all and then throw out the old fans, they aren't needed.

-Thermal paste is also not needed since the AIO will have pre applied thermal paste. Save purchasing your own tube for down the road when it comes time to clean and reapply.

 

If you decided to change anything, pcpartpicker has a great built in compatibility tool, and if you're ever confused about putting it together (wiring or anything), there are a million step by step videos on YouTube that can set you up in the right direction. Just don't watch the video from the Verge. Hope this helps!

 

edit: also welcome to the forum!

Thank you! Super helpful.

[aisle9]

Don't waste tons of money on Noctuas. They're nice, yes, but they're fans, and you'll get 90% of the performance and 90% of the silence out of a $30 box of 5 Arctic PWM 120mm fans off of Amazon. Put the money you'd be dropping into those (what, $30 per fan?) to get yourself a higher-tier NVIDIA card if machine learning is on the table, and a higher-wattage PSU if needed. The RMx series is a great choice.

[mpfiest]

32 minutes ago, RONOTHAN## said:

Please use PCPartPicker.com for making the parts list, it does the initial compatibility checks (it's not perfect, there are some parts that it says are compatible that actually aren't, but it's a good first check), finds the cheapest parts available, and is also much easier to view than a slew of Amazon links.

 

Anyway, this is probably the system I'd actually build: https://pcpartpicker.com/list/PfD97R

1. The cooler is a bit cheaper for about the same performance. The thermal paste is unnecessary as CPU coolers come with some.
2. The motherboard is more feature rich, though admittedly will require doing a BIOS update to work (this isn't too difficult, look up BIOS Flashback on MSI motherboards if you want to see a tutorial on how to do so)
3. The RAM is DDR5 as a) it's faster, especially for workstation tasks and b) it's what the motherboard supports (the list you have would not work as you have a DDR5 motherboard and a DDR4 memory kit)
4. The SSD is higher capacity, ever so slightly faster, and currently Western Digital has a better reputation than Samsung (Samsung has been plagued with QC issues over the past year or so). 
5. The case is a bit cheaper, though still super easy to build in for a first time builder. The Noctua fans you have aren't necessary.
6. The PSU is plenty good enough, a 1000W unit is unnecessary.
7. The GPU is a 4070 Ti. If you're planning on doing deep learning and other types of machine learning algorithms, you really should be using a Nvidia card. Most of the libraries out there for doing this are designed around using Nvidia technologies such as CUDA and the Tensor cores, and while there are some that can use AMD cards like the 6700 XT you selected, they're kind of rare and the Nvidia cards are generally better supported in this field. That said, if you want to save a bit of money, dropping down to the regular 4070 wouldn't lose you too much machine learning performance, and realistically for the games you're playing it should still be plenty adequate. 

There is a white theme to it, though that was mostly by accident. 

So I considered the tensor cores but the deep learning aspect of my career is likely a year away. 
if I were to go that route is there options that would fit my budget? And would it have a negative effect on gaming? If it wouldn’t go beyond budget and still works for gaming I would definitely prefer that.

 

for case cooling, I am a bit paranoid when it comes to overheating and am happy to pay a bit more for overkill. Does the case you provided come with fans / if I were to upgrade further what would you recommend?

 

As for the wiring, is that included in the PCPartPicker link you sent?

 

edit: if you were to beef up some aspect of the build to get closer to the 2K mark, what would you do?

[RONOTHAN##]

17 minutes ago, mpfiest said:

if I were to go that route is there options that would fit my budget? And would it have a negative effect on gaming? If it wouldn’t go beyond budget and still works for gaming I would definitely prefer that.

Yeah, you can get a Nvidia card in your budget, the 4070 Ti I mentioned. Technically you could do a little better for gaming with something like a 7900 XTX, you should be able to get that in the budget and it is a fair bit faster, though for the games you listed I wouldn't exactly call that a necessity. 

 

17 minutes ago, mpfiest said:

for case cooling, I am a bit paranoid when it comes to overheating and am happy to pay a bit more for overkill. Does the case you provided come with fans / if I were to upgrade further what would you recommend?

The case comes with two massive fans, one regular sized one for the back, and the CPU cooler comes with 3 more, so 6 in total. They aren't quite as good as the Noctuas, but they're far from bad. As for overheating, you're using a 13th gen Intel chip, those hit the thermal limits in the best of situations so there isn't really much you can do about it. That cooler should keep it from throttling too badly. 

 

17 minutes ago, mpfiest said:

As for the wiring, is that included in the PCPartPicker link you sent?

 

Cables come with the power supply, there's no need to buy cables separately or anything like that if that's what you're talking about. 

 

17 minutes ago, mpfiest said:

edit: if you were to beef up some aspect of the build to get closer to the 2K mark, what would you do?

Maybe upgrade to a 13900K/14900K, though that's well into the point of diminishing returns. What is more likely is that I'd get a second SSD, a 14700K instead (this is a bit faster than the 13700K thanks to the extra E cores), or just get a second monitor with the leftover budget. 

[brob]

Use pcpartpicker.com. If you have to purchase from a single merchant you can add that to the filter.

 

There is usually no need to get an NVMe drive with a heat sink. Quality motherboards include heatsinks for one or more NVMe m.2 drives.

 

Its unlikely six fans are necessary. Fans are needed to establish airflow through the case. Two intake and one or two exhaust are usually sufficient. Take into account fans that come with the case and AIO. The total fan count of your list is 12.

 

New CPU coolers come with thermal compound. There is generally no need to buy additional. Further, 8g of compound would do dozens of PC.

 

The PSU has far more capacity than needed. A 750w model would be quite sufficient.

 

I'd suggest an i7-14700K. It has more cores and higher performance than the i7-13700K.

 

The memory is NOT compatible with the motherboard. DDR5 is required.

 

PCPartPicker Part List

CPU: Intel Core i7-14700K 3.4 GHz 20-Core Processor  ($401.62 @ Newegg) 
CPU Cooler: Deepcool AK620 DIGITAL WH 68.99 CFM CPU Cooler  ($87.98 @ Newegg) 
Motherboard: Asus PRIME Z790-A WIFI ATX LGA1700 Motherboard  ($249.99 @ Amazon) 
Memory: G.Skill Ripjaws S5 64 GB (2 x 32 GB) DDR5-6000 CL30 Memory  ($212.31 @ MemoryC) 
Storage: Samsung 990 Pro 4 TB M.2-2280 PCIe 4.0 X4 NVME Solid State Drive  ($249.99 @ Adorama) 
Video Card: MSI MECH 2X V1 Radeon RX 6750 XT 12 GB Video Card  ($321.99 @ Newegg) 
Case: Lian Li LANCOOL 216 ATX Mid Tower Case  ($89.00 @ Newegg Sellers) 
Power Supply: Thermaltake Toughpower GF1 Snow - TT Premium 850 W 80+ Gold Certified Fully Modular ATX Power Supply  ($109.99 @ Amazon) 
Total: $1722.87
Prices include shipping, taxes, and discounts when available
Generated by PCPartPicker 2023-11-23 23:23 EST-0500

[igormp]

11 hours ago, mpfiest said:

XFX Speedster SWFT309 AMD Radeon RX 6700 XT

Since you're going to do ML and work with data, I'd avoid AMD GPUs and go for an Nvidia one instead due to CUDA.

11 hours ago, RONOTHAN## said:

The GPU is a 4070 Ti. If you're planning on doing deep learning and other types of machine learning algorithms, you really should be using a Nvidia card. Most of the libraries out there for doing this are designed around using Nvidia technologies such as CUDA and the Tensor cores, and while there are some that can use AMD cards like the 6700 XT you selected, they're kind of rare and the Nvidia cards are generally better supported in this field. That said, if you want to save a bit of money, dropping down to the regular 4070 wouldn't lose you too much machine learning performance, and realistically for the games you're playing it should still be plenty adequate. 

FWIW, 4060ti 16gb would be way better than the 4070 and 4070ti for ML, but worse for games.

[mpfiest]

On 11/23/2023 at 11:23 PM, RONOTHAN## said:

Yeah, you can get a Nvidia card in your budget, the 4070 Ti I mentioned. Technically you could do a little better for gaming with something like a 7900 XTX, you should be able to get that in the budget and it is a fair bit faster, though for the games you listed I wouldn't exactly call that a necessity. 

 

The case comes with two massive fans, one regular sized one for the back, and the CPU cooler comes with 3 more, so 6 in total. They aren't quite as good as the Noctuas, but they're far from bad. As for overheating, you're using a 13th gen Intel chip, those hit the thermal limits in the best of situations so there isn't really much you can do about it. That cooler should keep it from throttling too badly. 

 

Cables come with the power supply, there's no need to buy cables separately or anything like that if that's what you're talking about. 

 

Maybe upgrade to a 13900K/14900K, though that's well into the point of diminishing returns. What is more likely is that I'd get a second SSD, a 14700K instead (this is a bit faster than the 13700K thanks to the extra E cores), or just get a second monitor with the leftover budget. 

I built the PC you recommended and it is awesome! I upgraded to a 14700K, I'm so happy with the result, and it looks awesome. Super grateful for your help! Thank you.