From WikiChip
Difference between revisions of "amd/ryzen threadripper/3970x"
< amd‎ | ryzen threadripper

(Official listed max memory for the Threadripper 3970X is 512 GB. CPU only supports UDIMM with a current max capacity of 32 GB DIMMs. 256GB being the current "limit", but 512GB is the absolute limit.)
Line 68: Line 68:
 
|type=DDR4-3200
 
|type=DDR4-3200
 
|ecc=Yes
 
|ecc=Yes
|max mem=512 GB
+
|max mem=256 GB
 
|controllers=4
 
|controllers=4
 
|channels=4
 
|channels=4

Revision as of 10:39, 29 March 2020

Edit Values
Ryzen Threadripper 3970X
General Info
DesignerAMD
ManufacturerTSMC, GlobalFoundries
Model Number3970X
Part Number100-000000011
MarketDesktop
IntroductionNovember 7, 2019 (announced)
November 25, 2019 (launched)
Release Price$1,999.00
ShopAmazon
General Specs
FamilyRyzen Threadripper
Series3900
LockedNo
Frequency3,700 MHz
Turbo Frequency4,500 MHz
Bus rate8 × 16 GT/s
Clock multiplier37
Microarchitecture
ISAx86-64 (x86)
MicroarchitectureZen 2
ChipsetTRX40
Core NameCastle Peak
Process7 nm, 12 nm
Transistors23,540,000,000
TechnologyCMOS
MCPYes (5 dies)
Word Size64 bit
Cores32
Threads64
Multiprocessing
Max SMP1-Way (Uniprocessor)
Electrical
TDP280 W
Tjunction0 °C – 68 °C
Packaging
Unknown package "amd,socket_trx4"
Succession

Ryzen Threadripper 3970X is a 64-bit dotriaconta-core high-performance x86 desktop microprocessor introduced by AMD in November 25, 2019. The 3970X, which is based on their Zen 2 microarchitecture, is fabricated on TSMC's 7 nm process. The 3970X operates at a base frequency of 3.7 GHz with a TDP of 280 W and a boost of up to 4.5 GHz. This MPU supports up to ? TiB of quad-channel DDR4-3200 memory.


Cache

Main article: Zen 2 § Cache

[Edit/Modify Cache Info]

hierarchy icon.svg
Cache Organization
Cache is a hardware component containing a relatively small and extremely fast memory designed to speed up the performance of a CPU by preparing ahead of time the data it needs to read from a relatively slower medium such as main memory.

The organization and amount of cache can have a large impact on the performance, power consumption, die size, and consequently cost of the IC.

Cache is specified by its size, number of sets, associativity, block size, sub-block size, and fetch and write-back policies.

Note: All units are in kibibytes and mebibytes.
L1$2 MiB
2,048 KiB
2,097,152 B
L1I$1 MiB
1,024 KiB
1,048,576 B
32x32 KiB8-way set associative 
L1D$1 MiB
1,024 KiB
1,048,576 B
32x32 KiB8-way set associative 

L2$16 MiB
16,384 KiB
16,777,216 B
0.0156 GiB
  32x512 KiB8-way set associativewrite-back

L3$128 MiB
131,072 KiB
134,217,728 B
0.125 GiB
  8x16 MiB  

Memory controller

[Edit/Modify Memory Info]

ram icons.svg
Integrated Memory Controller
Max TypeDDR4-3200
Supports ECCYes
Max Mem256 GB
Controllers4
Channels4
Max Bandwidth95.37 GiB/s
97,658.88 MiB/s
102.403 GB/s
102,402.758 MB/s
0.0931 TiB/s
0.102 TB/s
Bandwidth
Single 23.84 GiB/s
Double 47.68 GiB/s
Quad 95.37 GiB/s

Expansions

[Edit/Modify Expansions Info]

ide icon.svg
Expansion Options
PCIeRevision: 4.0
Max Lanes: 64
Configuration: 3x16+4x4


Features

[Edit/Modify Supported Features]

Cog-icon-grey.svg
Supported x86 Extensions & Processor Features
MMXMMX Extension
EMMXExtended MMX Extension
SSEStreaming SIMD Extensions
SSE2Streaming SIMD Extensions 2
SSE3Streaming SIMD Extensions 3
SSSE3Supplemental SSE3
SSE4.1Streaming SIMD Extensions 4.1
SSE4.2Streaming SIMD Extensions 4.2
SSE4aStreaming SIMD Extensions 4a
AVXAdvanced Vector Extensions
AVX2Advanced Vector Extensions 2
ABMAdvanced Bit Manipulation
BMI1Bit Manipulation Instruction Set 1
BMI2Bit Manipulation Instruction Set 2
FMA33-Operand Fused-Multiply-Add
AESAES Encryption Instructions
RdRandHardware RNG
SHASHA Extensions
ADXMulti-Precision Add-Carry
CLMULCarry-less Multiplication Extension
F16C16-bit Floating Point Conversion
x86-1616-bit x86
x86-3232-bit x86
x86-6464-bit x86
RealReal Mode
ProtectedProtected Mode
SMMSystem Management Mode
FPUIntegrated x87 FPU
NXNo-eXecute
SMTSimultaneous Multithreading
AMD-ViAMD-Vi (I/O MMU virtualization)
AMD-VAMD Virtualization
TSMETransparent SME
SenseMISenseMI Technology
Boost 2Precision Boost 2
Has subobject
"Has subobject" is a predefined property representing a container construct and is provided by Semantic MediaWiki.
Ryzen Threadripper 3970X - AMD#pcie +
base frequency3,700 MHz (3.7 GHz, 3,700,000 kHz) +
bus links8 +
bus rate16,000 MT/s (16 GT/s, 16,000,000 kT/s) +
chipsetTRX40 +
clock multiplier37 +
core count32 +
core nameCastle Peak +
designerAMD +
die count5 +
familyRyzen Threadripper +
first announcedNovember 7, 2019 +
first launchedNovember 25, 2019 +
full page nameamd/ryzen threadripper/3970x +
has advanced vector extensionstrue +
has advanced vector extensions 2true +
has amd amd-v technologytrue +
has amd amd-vi technologytrue +
has amd precision boost 2true +
has amd sensemi technologytrue +
has amd transparent secure memory encryption technologytrue +
has ecc memory supporttrue +
has featureAdvanced Vector Extensions +, Advanced Vector Extensions 2 +, Advanced Encryption Standard Instruction Set Extension +, SenseMI Technology + and Precision Boost 2 +
has locked clock multiplierfalse +
has simultaneous multithreadingtrue +
has x86 advanced encryption standard instruction set extensiontrue +
instance ofmicroprocessor +
is multi-chip packagetrue +
isax86-64 +
isa familyx86 +
l1$ size2,048 KiB (2,097,152 B, 2 MiB) +
l1d$ description8-way set associative +
l1d$ size1,024 KiB (1,048,576 B, 1 MiB) +
l1i$ description8-way set associative +
l1i$ size1,024 KiB (1,048,576 B, 1 MiB) +
l2$ description8-way set associative +
l2$ size16 MiB (16,384 KiB, 16,777,216 B, 0.0156 GiB) +
l3$ size128 MiB (131,072 KiB, 134,217,728 B, 0.125 GiB) +
ldate3000 +
manufacturerTSMC + and GlobalFoundries +
market segmentDesktop +
max cpu count1 +
max junction temperature341.15 K (68 °C, 154.4 °F, 614.07 °R) +
max memory bandwidth95.37 GiB/s (97,658.88 MiB/s, 102.403 GB/s, 102,402.758 MB/s, 0.0931 TiB/s, 0.102 TB/s) +
max memory channels4 +
microarchitectureZen 2 +
min junction temperature273.15 K (0 °C, 32 °F, 491.67 °R) +
model number3970X +
nameRyzen Threadripper 3970X +
part number100-000000011 +
process7 nm (0.007 μm, 7.0e-6 mm) + and 12 nm (0.012 μm, 1.2e-5 mm) +
release price$ 1,999.00 (€ 1,799.10, £ 1,619.19, ¥ 206,556.67) +
series3900 +
smp max ways1 +
supported memory typeDDR4-3200 +
tdp280 W (280,000 mW, 0.375 hp, 0.28 kW) +
technologyCMOS +
thread count64 +
transistor count23,540,000,000 +
turbo frequency4,500 MHz (4.5 GHz, 4,500,000 kHz) +
word size64 bit (8 octets, 16 nibbles) +