Heterogeneous V2X real-time SM2 authentication method, system and equipment based on GPU acceleration
By adopting GPU-accelerated heterogeneous V2X real-time SM2 authentication method in V2X communication, combining resource isolation and multi-level parallel optimization of CPU and GPU, the balance problem of delay and throughput in V2X systems is solved, and efficient signature and verification of safety-critical and non-safety-critical applications is achieved, and computing performance is improved.
Patent Information
- Application Number
- CN202510747197.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-05
- Publication Date
- 2025-08-01
AI Technical Summary
The prior art is difficult to effectively balance the latency and throughput requirements of safety-critical and non-safety-critical applications in V2X communication, especially in the acceleration and authentication of SM2 elliptic curve signature algorithms in the Internet of Vehicles, and cannot meet the requirements of low latency and high throughput at the same time.
The heterogeneous V2X real-time SM2 authentication method based on GPU acceleration is adopted. By processing SM3 password hash function on the CPU side, and batch SM2 elliptic curve signature generation and verification are carried out on the GPU side. Combined with CUDA Kernel Launch Parameter Auto Tuning technology, resource isolation and multi-level parallel optimization are realized, and safety-critical and non-safety-critical applications are handled respectively.
On the premise of ensuring security and real-time, it effectively takes into account the needs of low latency and high throughput, providing an efficient signature and signature verification solution for V2X systems, improving computing performance, and meeting application needs in different scenarios.
Smart Images
Figure CN120415744A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to multiple fields such as intelligent driving, high-performance computing, and cryptography, and particularly relates to a heterogeneous V2X real-time SM2 authentication method, system, and device based on GPU acceleration. Specifically, on the vehicle networking application layer, the SM2 elliptic curve algorithm is used to provide message integrity and non-repudiation, and the CUDA Kernel Launch Parameter AutoTuning technology is introduced to significantly improve the cryptographic operation performance, so as to meet the different performance requirements of two types of applications, namely security-critical and non-security-critical applications, under the V2X application layer. Background Art
[0002] In recent years, the GPU has become an indispensable multi-core accelerator in the field of general computing, and its hardware design is optimized for supporting large-scale parallel computing. Modern GPUs usually integrate tens of thousands of streamlined computing cores (such as CUDA cores), which are oriented towards large-scale parallel operations, rather than emphasizing the complexity or versatility of a single core. There is also a global memory of several GB to tens of GB level, and all thread blocks can access this memory space; through the memory access coalescing technology, the actual bandwidth of the global memory is significantly improved.
[0003] In addition, the GPU is also equipped with low-latency and high-bandwidth shared memory inside each thread block for efficient data exchange between threads, usually with a capacity of several tens of KB. To accelerate the access to read-only data, the GPU also provides on-chip constant memory and texture memory, both of which can further improve the access efficiency of frequently used data.
[0004] In terms of execution scheduling, the GPU divides the computing resources into warps composed of 32 consecutive threads. The threads within each warp synchronously execute the same instructions in the same scheduling unit, reflecting the characteristics of the single instruction multiple threads (SIMT) architecture. This design is extremely suitable for processing a large number of repetitive computing tasks, enabling large-scale data to be efficiently processed in parallel. Due to the simplified single-core design and low power consumption, the GPU can integrate a large number of such cores, thus completing a huge amount of parallel computing work in a short time. At the same time, the multi-level memory hierarchy (global memory, shared memory, constant / texture memory) works together to provide a solid hardware support for high-throughput parallel computing, effectively improving the overall computing performance and resource utilization With the rapid evolution of intelligent transportation systems, the data exchange between infrastructure and vehicles in vehicle-infrastructure cooperative communications is growing exponentially. Soto et al. focus on two typical applications based on C-V2X technology: one centered on road safety, and the other on traffic efficiency and operational optimization, and elaborate on each in their paper. Xiong et al. categorize V2X services into road safety (such as collision warning and intersection warning) and operational efficiency / information services (such as real-time traffic updates, weather information push, and in-vehicle payment), emphasizing the different latency and reliability requirements of these two types of applications. Based on this, the requirements for different security levels within the V2X communication architecture can be divided into two categories: "safety-critical" and "non-safety-critical." Safety-critical applications directly impact the safety of vehicles, drivers, and pedestrians, requiring extremely low latency (milliseconds), high reliability, and system stability, and strong authentication and encryption mechanisms. Typical scenarios include cooperative driving, platooning, collision warning, and emergency braking. In contrast, non-safety-critical applications have relatively relaxed security requirements, but still need to ensure the integrity and reliability of massive data transmission. They mainly cover areas such as traffic information services, environmental data sharing, and in-vehicle connectivity.
[0005] This invention applies heterogeneous computing acceleration to information reliability authentication at the application layer of V2X. It implements and evaluates the signature generation and verification performance of SM2-based elliptic curve digital signatures in different V2X scenarios, including CPU-side SM3 cryptographic hash functions and GPU-side batch acceleration operations. For safety-critical scenarios, a multi-threaded instance parallelization strategy is proposed to reduce latency; in non-safety-critical scenarios, throughput is improved by increasing batch size. Experimental results show that both solutions achieve an effective balance between latency and throughput, providing a viable SM2 signature technology path for real-time secure V2X communications.
[0006] Before signing, the system first calls the SM3 cryptographic hash function algorithm on the CPU to concatenate the user's identity (ID) with the message to be signed, M, to generate the message digest e. The system implements batch signature generation and verification using the SM2 algorithm on the GPU: After receiving several signature requests, the CPU caches them and periodically sends them to the GPU thread group, triggering parallel execution.
[0007] This paper introduces a non-safety-critical application sublayer in the application layer of the WAVE protocol stack to handle non-safety-critical messages that require security protection. At the same time, for the two types of message sublayers of safety-critical and non-safety-critical, kernel isolation and resource partitioning technologies are adopted to implement an independent SM2 signature and verification operation environment. By completing the SM3 cryptographic hash function at the CPU side and batch accelerating the generation and verification of SM2 elliptic curve signatures at the GPU side, performance optimization is achieved in low-latency and high-throughput scenarios respectively, providing a feasible technical solution for V2X secure communication. To avoid interference introduced by non-safety-critical tasks during signature and verification, the system uses a separation kernel to ensure that the execution of safety-critical tasks is not affected by non-safety-critical tasks. Summary of the Invention
[0008] Aiming at the deficiencies of the prior art, the present invention provides a heterogeneous V2X real-time SM2 authentication method, system and device based on GPU acceleration. Specifically, in the field of V2X vehicle networking, the application of the real-time SM2 elliptic curve signature algorithm based on heterogeneous computing acceleration decomposes the core operation module of the SM2 elliptic curve digital signature algorithm at the thread level with fine granularity, and combines a multi-level parallel optimization strategy, and at the same time introduces the CUDA Kernel Launch Parameter Auto Tuning technology to significantly improve the operation performance of elliptic curve signature generation and verification in the V2X vehicle networking application layer.
[0009] The implementation of the present invention is divided into two parts: one part is to statically partition the kernel resources: allocate 22-core CPU / GPU computing units for the safety-critical application layer and the non-safety-critical application layer respectively to achieve equal resource isolation and load balancing. The SM2 signature and verification modules are independently deployed within each isolation domain to avoid core competition and cache interference, and improve the latency controllability and throughput stability of each.
[0010] First, the statically partitioned CPU splices the user's identity (ID) and the original message M through the SM3 cryptographic hash function to calculate a unique message digest e. The signer uses the GPU to batch process and generate a signature pair (r, s) through the SM2 elliptic curve digital signature algorithm according to its private key and the digest message e; the verifier verifies the signature pair (r, s), the original message M and the signer's public key through GPU processing for signature verification.
[0011] In applications with extremely high real-time requirements such as V2X, after successful signature verification, in addition to the above steps, it is also necessary to strictly check multi-dimensional elements such as the timeliness of the signature (signature timestamp), context legality (message source matching the business scenario), and certificate status (revocation list or Online Certificate Status Protocol), and make joint decisions in combination with the vehicle's real-time status to ensure rapid response and handling in emergency situations. In applications such as general traffic information services or navigation assistance, after passing signature verification, the focus is mainly on message content interpretation and lightweight context checking, and then the message is handed over to the corresponding business logic module for processing without overly consuming computing resources for multiple security validations.
[0012] The SM2 elliptic curve digital signature algorithm combines the identity binding mechanism for the digest with the SM3 cryptographic hash function, as well as the security based on the elliptic curve discrete logarithm problem, providing an efficient and secure digital signature solution for application scenarios such as the Internet of Vehicles that require high reliability and high non-repudiation. By flexibly configuring the timeliness and context checking in the signature and signature verification processes, it can support real-time communication requirements from resource-constrained lightweight applications to security-critical ones.
[0013] Using the CUDA Kernel Launch Parameter Auto Tuning technology, first conduct experiments on TPI (the number of threads set for each instance) to evaluate the performance of the fast algorithm. After a large number of experiments, select several most suitable TPIs, and with the TPI fixed, dynamically change the number of instances each time. Through the CUDA Kernel Launch Parameter Auto Tuning method, for different TPIs, change the number of instances each time for different TPIs, and record the latency and throughput (speed) for analysis to obtain the result of the optimal number of instances.
[0014] Another part is to complete the application of the real-time SM2 elliptic curve signature algorithm based on heterogeneous computing for V2X security-critical scenarios and V2X non-security-critical scenarios. The framework diagram of the system is as Figure 1 shown.
[0015] The present invention also provides a heterogeneous V2X real-time SM2 authentication system based on GPU acceleration, which embeds the method described in the present invention.
[0016] The present invention also provides a heterogeneous V2X real-time SM2 authentication device based on GPU acceleration, which embeds the method described in the present invention.
[0017] The beneficial effects of the present invention are as follows: By incubating a non-safety-critical application sub-layer in the WAVE application layer, combining SM2 / SM3 algorithms, and applying static core partitioning technology, this paper has realized an independent security processing framework for safety-critical and non-safety-critical messages in the V2X system. On the premise of ensuring security and real-time performance, it effectively takes into account the requirements of two types of scenarios: low latency and high throughput, providing a practical reference for the large-scale deployment of the SM2 signature technology in the vehicle networking. Description of the Drawings
[0018] Figure 1 It is a framework diagram for introducing a non-critical security application layer based on the application layer of the WAVE protocol stack.
[0019] Figure 2 It is a comparison chart of the SM2 elliptic curve signature generation speed and latency based on heterogeneous computing acceleration in safety-critical scenarios.
[0020] Figure 3 It is a comparison of the SM2 elliptic curve signature generation speed based on heterogeneous computing acceleration with other hardware accelerations in safety-critical scenarios.
[0021] Figure 4 It is a comparison of the SM2 elliptic curve signature latency based on heterogeneous computing acceleration with other hardware accelerations in safety-critical scenarios.
[0022] Figure 5 It is a comparison chart of the SM2 elliptic curve signature verification speed and latency based on heterogeneous computing acceleration in safety-critical scenarios.
[0023] Figure 6 It is a comparison of the SM2 elliptic curve signature verification speed based on heterogeneous computing acceleration with other hardware accelerations in safety-critical scenarios.
[0024] Figure 7 It is a comparison of the SM2 elliptic curve signature verification latency based on heterogeneous computing acceleration with other hardware accelerations in safety-critical scenarios.
[0025] Figure 8 It is a comparison chart of the SM2 elliptic curve signature generation speed and latency based on heterogeneous computing acceleration in non-safety-critical scenarios.
[0026] Figure 9 It is a comparison of the SM2 elliptic curve signature generation speed based on heterogeneous computing acceleration with other hardware accelerations in non-safety-critical scenarios.
[0027] Figure 10 It is a comparison of the SM2 elliptic curve signature latency based on heterogeneous computing acceleration with other hardware accelerations in non-safety-critical scenarios.
[0028] Figure 11 It is a comparison chart of the SM2 elliptic curve signature verification speed and latency based on heterogeneous computing acceleration in non-safety-critical scenarios.
[0029] Figure 12 For accelerating the speed of SM2 elliptic curve signature verification based on heterogeneous computing in non-safety-critical scenarios, comparison with the safety-critical scenario of the present invention and other hardware accelerations.
[0030] Figure 13 For comparing the latency of accelerating SM2 elliptic curve signature verification based on heterogeneous computing in non-safety-critical scenarios with the safety-critical scenario of the present invention and other hardware accelerations. Detailed implementation manners
[0031] The present invention will be described in detail below with reference to the accompanying drawings, and the objectives and effects of the present invention will become more obvious.
[0032] Embodiment The CPU adopted is AMD EPYC 7K62 48-Core Processor, and the GPU used is NVIDIA GeForce RTX 4090. The server runs the Ubuntu 20.04.6 system, and the NVCC used for the GPU version is 12.3. The reason for selecting NVIDIA GeForce RTX 4090 is that the integer operation performance of DRIVE Thor released by NVIDIA at GTC in 2022 is 1000 TOPS (INT8) in non-sparse cases, far exceeding 660.6 TOPS (INT8) of NVIDIA GeForce RTX 4090. The server CPU cores are allocated, and both the safety-critical application layer and the non-safety-critical application layer have 22 CPU cores, and one NVIDIA GeForce RTX 4090 GPU is allocated to generate and verify SM2 elliptic curve signatures.
[0033] The specific implementation process of this method is as follows: Step 1: Through the processing of the SM3 cryptographic hash function on the CPU side, the user identity (ID) and the original message M are concatenated to generate a message digest e. Different from other elliptic curve signatures, in the process of generating the message digest e for SM2 elliptic curve digital signatures, its own parameters and the user identity (user ID) are used. The calculation formula of the message digest e is as follows:
[0034] Among them, is the binding value of the user identifier and the public key, and M is the original message. And The calculation formula of is as follows:
[0035] In the above formula, SM3 is Cryptographic hash function is the bit length of the ID. is the user identity identifier. The variables a and b are two parameters defining the shape of the SM2 curve. and are the base point coordinates of the SM2 elliptic curve, and and are the user's public key coordinates. The Weierstrass standard form equation of the SM2 elliptic curve is as follows:
[0036] Among them, the values of a and b are shown as follows:
[0037]
[0038] For the remaining parameters and detailed processes, please refer to GB / T 32918.2-2016.
[0039] The following is an example of the corresponding information. All the parameters in the examples in this patent are generated based on M and the user ID in the examples, and corresponding verification can be carried out.
[0040] This invention draws on the original message M format defined in ASN.1 in ETSI ITS-G5 in Europe and gives the following example: CAM ::= SEQUENCE { header CamHeader, camParameters CamParameters } CamHeader { protocolVersion = 1, messageID = 2, -- CAM message stationID = 0x12345678 } CamParameters { basicContainer { stationType = 5, -- Passenger car referencePosition { latitude = 486520000, -- 48.6520000°N longitude = 92842000, -- 9.2842000°E positionConfidence = {...} } }, highFrequencyContainer { basicVehicleContainerHighFrequency { heading = 9000, -- 90.00° speed = 1389, -- 13.89 m / s driveDirection = 0 -- forward } } } The text of the original message M is converted to UTF-8 encoding, and the user ID is set to the name of the corresponding intelligent driving user, which is set here as:
[0041] Take the private key as:
[0042] Step 2: Use NVIDIA's CGBN library to achieve batch acceleration of SM2 elliptic curve signature generation and verification on the GPU side. Specifically: The signer uses its private key and the digest message e through the SM2 elliptic curve digital signature algorithm, and uses the GPU to batch process to generate the signature pair (r, s). The verifier uses the GPU signature pair (r, s), the original message M, and the signer's public key to verify the signature.
[0043] Step 3: Use the message digest e generated in batches on the CPU side, and use the GPU for batch SM2 elliptic curve signature generation and verification, where the random number k used in the elliptic curve signature process is generated on the GPU.
[0044] This invention takes the original message M and the user ID: Qitao Feng as an example, and the corresponding message digest e generated is:
[0045] This invention uses the GPU to perform SM2 elliptic curve digital signature on the message digest e. For the convenience of verification, it is assumed here that the random number k is:
[0046] The results obtained by the GPU signature of the present invention include: (r, s) of the signature part and the abscissa of the generated public key and the ordinate .
[0047] Based on the given information: user identity ID, private key , digest message e, random number k, using the GPU for elliptic curve signature, the generated signature information is as follows:
[0048]
[0049]
[0050]
[0051]
[0052] Among them, the above signature information can be fully brought in for verification. By bringing in the corresponding parameters for signature verification, it is found that the verification is successful, proving the non-repudiation and integrity of the information.
[0053] In the verification stage, the generation of the message digest e is similar to that in step 1 of the SM2 elliptic curve digital signature. Using the generated public key , the original message M, the bit length of the identification ID SM2 elliptic curve parameters a, b and perform the SM3 cryptographic hash function algorithm to generate the message digest e. To verify the non-repudiation and integrity of the information, we need to calculate the intermediate variable t
[0054] where r and s are the signature parts and n is the order of the SM2 elliptic curve base point. Then calculate the elliptic curve point
[0055] where G is the base point coordinate of the SM2 elliptic curve, s corresponds to the signature part, t is the intermediate variable calculated previously is the public key. Then calculate
[0056] Finally, judge whether R is equal to r in the signature content. If they are equal, it means that the information transmission is error-free and there is no possibility of being tampered with, that is, the non-repudiation and integrity in the information transmission process are verified.
[0057] Step 4: Test the overall integrity of heterogeneous computing for SM2 elliptic curve signature generation and verification in the vehicle networking scenario (V2X).
[0058] Step 5: Apply the test results to the sub-layer of wave in the vehicle networking scenario, and introduce a non-safety-critical application layer based on the original safety-critical application layer.
[0059] Step 6: Perform static partitioning of CPU kernel resources by configuring the Linux kernel, and allocate 22-core CPU computing units for the safety-critical application layer and the non-safety-critical application layer respectively. Both application layers are equipped with GPUs, which are responsible for processing elliptic curve signature and verification in their respective application layers, so as to avoid core competition and cache interference, and improve the latency controllability and throughput stability of each layer.
[0060] Step 7: Under the condition of the same number of controlled kernels, test the generation and verification of SM2 elliptic curve digital signatures in different parameter environments by dynamically adjusting the number of threads responsible for each instance and dynamically adjusting the number of instances copied from the CPU to the GPU at one time, that is, the CUDA Kernel Launch Parameter AutoTuning technology.
[0061] Step 8: Select the openssl library to conduct experiments on elliptic curve signature and verification based on the CPU, and compare the performance with the results of a paper on accelerating the SM2 elliptic curve signature algorithm by FPGA in the V2X field last year. Finally, the discovered results have obvious advantages.
[0062] The generation and verification of SM2 elliptic curve signatures based on V2X are implemented on the GPU, and through the CUDA Kernel Launch Parameter Auto Tuning technology, the configuration and application of different parameters in safety-critical scenarios and non-safety-critical scenarios are realized. In the V2X safety-critical scenario, the design and parameter configuration achieve 683342 sign / s for SM2 signature generation and 566608 verify / s for SM2 signature verification. In the V2X non-safety-critical scenario, the design and parameter configuration achieve 2281706 sign / s for SM2 signature generation and 949877 verify / s for SM2 signature verification.
[0063] Final result comparison of the experimental data of the present invention In the safety-critical scenario, to meet the requirements of the V2X safety-critical scenario, extremely low latency and relatively appropriate throughput are required to meet the needs of the V2X application layer. Among them, the latency and speed of SM2 elliptic curve signature generation in the safety-critical scenario are as Figure 2As shown in the figure. When the number of instances is selected as 2048, it takes 1.5 ms to generate a signature once, and 2048 signatures are generated. While achieving the millisecond-level signature generation process, it also realizes batch generation. The speed comparison of SM2 elliptic curve digital signature generation is shown in Table 2. It is more than 89 times faster than LLP-ECCA based on FPGA, and the throughput should be the highest at present. It supports generating 1024 signatures within 1.5 ms, and this efficiency is quite high. Then, the comparison of the situation of batch executing a signature once with FPGA and CPU is as Figure 3 , Figure 4 shown. The time required to generate an average signature has even entered the microsecond level, and it can be flexibly adapted. In the process of taking care of efficiency, sacrificing a little latency for batch generation is also a trade-off, and it can meet the conditions and requirements in V2X safety-critical scenarios.
[0064] During the verification process of elliptic curve digital signature, when TPI = 16, the effect is still the best. For different TPIs and different numbers of instances, the curve graph of signature verification latency is as Figure 5 shown. The optimal number of instances processed each time is obtained as 2048, and the optimal TPI = 16. The latency (wall time) of elliptic curve signature verification for 2048 instances at a time is only 3.6 ms. In the case where the number of instances is 2048 and TPI = 16, the speed comparison of SM2 elliptic curve digital signature verification is as Figure 6 shown. The process of signature verification designed is for 2048 instances at a time, and it only takes 3.6 ms. The time required to verify an average signature has even entered the microsecond level. Sacrificing latency for batch verification is a trade-off. The comparison of the latency of SM2 elliptic curve digital signature verification is as Figure 7 shown.
[0065] In V2X non-safety-critical scenarios, the latency can be appropriately increased, which can improve the throughput to meet the throughput requirements of non-safety-critical scenarios.
[0066] The scheme of selecting TPI = 16 and processing 16384 instances each time makes the latency of SM2 elliptic curve signature generation 23.8 ms, which fully meets the requirements of V2X non-safety-critical scenarios, and the throughput is 3.3 times that of V2X safety-critical scenarios. This is undoubtedly the best for non-safety scenarios that require higher throughput, and this situation has almost reached the bottleneck of the GPU. The speed comparison and latency comparison of SM2 elliptic curve signature generation in the non-critical safety layer are as Figure 8 shown. The comparison with the V2X safety-critical scenario designed by oneself and FPGA is as Figure 9, Figure 10 as shown
[0067] In the elliptic curve signature verification process, the case of TPI = 16 and the number of instances being 16384 was selected. The time required for elliptic curve signature verification in this scheme is 17.24 ms, which fully meets the V2X communication requirements in non-safety-critical scenarios. Moreover, the signature verification speed in non-safety scenarios is 1.67 times that in safety scenarios, and the throughput has been greatly improved. The speed comparison and latency comparison of SM2 elliptic curve signature verification in the non-critical safety layer are as Figure 10 shown. The comparison with the V2X safety-critical scenario designed by oneself and the FPGA is as Figure 12 , Figure 13 shown
[0068] The above content is a further detailed description of the present invention in combination with specific / preferred embodiments. It cannot be determined that the specific implementation of the present invention is only limited to these descriptions. For those of ordinary skill in the technical field to which the present invention pertains, without departing from the concept of the present invention, they can also make several substitutions or modifications to these described embodiments, and these substitution or modification methods should all be regarded as belonging to the protection scope of the present invention.
[0069] The parts not detailed in the present invention belong to the well-known technologies in the art.
Claims
1. A heterogeneous V2X real-time SM2 authentication method based on GPU acceleration, characterized in that It includes the following steps: Step 1: After processing by the SM3 cryptographic hash function on the CPU side, the user identity ID and the original message M are concatenated to generate a message digest e; Step 2: Utilize NVIDIA's CGBN library to achieve batch acceleration of SM2 elliptic curve signature generation and verification on the GPU side; Step 3: Use the message digest e batch-generated on the CPU side, and utilize the GPU to perform batch SM2 elliptic curve signature generation and verification, where the random number k used in the elliptic curve signature process is generated on the GPU; Step 4: In the vehicle networking scenario, test the overall integrity of the heterogeneous computing for SM2 elliptic curve signature generation and verification; Step 5: Apply the test results to the sub-layer of wave in the vehicle networking scenario, and introduce a non-security-critical application layer based on the original security-critical application layer; Step 6: Through configuring the Linux kernel, perform static partitioning of the CPU core resources, allocate 22-core CPU computing units for the security-critical application layer and the non-security-critical application layer respectively, and both application layers are provided with GPUs, which are respectively responsible for processing elliptic curve signature and verification in their respective application layers; Step 7: Under the condition of the same number of controlled cores, perform tests on the generation and verification of SM2 elliptic curve digital signatures in different parameter environments by dynamically adjusting the number of threads responsible for each instance and dynamically adjusting the number of instances copied from the CPU to the GPU at one time.
2. The heterogeneous V2X real-time SM2 authentication method based on GPU acceleration according to claim 1, wherein In Step 1, during the process of the SM3 cryptographic hash function algorithm, the original message M and the user identity ID need to be provided; the text of the original message M is converted to UTF-8 encoding, and the user identity ID is set as the name of the corresponding intelligent driving user.
3. The heterogeneous V2X real-time SM2 authentication method based on GPU acceleration according to claim 2, wherein, Step 2 is specifically implemented as follows: The signer, based on its private key and the digest message e, uses the SM2 elliptic curve digital signature algorithm to generate a signature pair (r, s) through batch processing by the GPU; the verifier processes the signature pair (r, s), the original message M, and the signer's public key through the GPU to perform signature verification.
4. The heterogeneous V2X real-time SM2 authentication method based on GPU acceleration according to claim 2, wherein The specific implementation of Step 3 is as follows: Perform SM2 elliptic curve digital signature on the message digest e through the GPU; The results obtained from GPU signature include: (r, s) of the signature part and the abscissa of the generated public key and the ordinate ; Based on the given information, using the GPU to perform elliptic curve signature, the generated signature information is as follows: , , , , ; The given information includes: user identity ID, private key , digest message e, random number k; Completely bring in the signature information for signature verification to prove the non-repudiation and integrity of the information.
5. Heterogeneous V2X real-time SM2 authentication system based on GPU acceleration, characterized in that, The method described in Claim 1 is embedded in this system.
6. Heterogeneous V2X real-time SM2 authentication device based on GPU acceleration, characterized in that, The method described in Claim 1 is embedded in this device.